Addressing concentrations of disadvantage in urban

0 downloads 0 Views 2MB Size Report
Table 7: Average distance from CBD to disadvantaged suburbs (kilometre) ............. ..... redevelopment projects such as those at Bonnyrigg (NSW) and Kensington ...... deindustrialised areas of cities such as Baltimore, Belfast, Buffalo, Detroit, ...
Addressing concentrations of disadvantage in urban Australia authored by

Hal Pawson, Kath Hulse and Lynda Cheshire for the

Australian Housing and Urban Research Institute at The University of New South Wales at Swinburne University of Technology

September 2015 AHURI Final Report No. 247 ISSN: 1834-7223 ISBN: 978-1-922075-97-0

Authors

Pawson, Hal

The University of New South Wales

Hulse, Kath

Swinburne University of Technology

Cheshire, Lynda

University of Queensland

Title

Addressing concentrations of disadvantage in urban Australia

ISBN

978-1-922075-97-0

Format

PDF

Key words

Spatial disadvantage, social disadvantage, residential mobility, social exclusion, social deprivation, low-income housing, housing choice.

Editor

Anne Badenhorst

Publisher

Australian Housing and Urban Research Institute Melbourne, Australia

Series

AHURI Final Report; no. 247

ISSN

1834-7223

Preferred citation

Pawson, H., Hulse, K. and Cheshire, L. (2015) Addressing concentrations of disadvantage in urban Australia, AHURI Final Report No.247. Melbourne: Australian Housing and Urban Research Institute. Available from: . [Add the date that you accessed this report: DD MM YYYY].

AHURI National Office

i

ACKNOWLEDGEMENTS This material was produced with funding from the Australian Government and the Australian state and territory governments. AHURI Limited gratefully acknowledges the financial and other support it has received from these governments, without which this work would not have been possible. AHURI comprises a network of university Research Centres across Australia. Research Centre contributions, both financial and in-kind, have made the completion of this report possible. This report is draws on a research program comprising a succession of linked studies, and involving numerous academic colleagues in addition to the authors of this report itself. From University of New South Wales (UNSW), these have included Dr Hazel Easthope, Dr Shanaka Herath and Andrew Tice. From Swinburne University of Technology, they have included Professor Terry Burke, Dr Wendy Stone, Dr Lucy Groenhart and Margaret Reynolds. From University of Queensland they have included Professor Andrew Jones and Dr Jonathan Corcoran. We also acknowledge the seminal contributions of Professor Bill Randolph and Associate Professor Simon Pinnegar (UNSW) as lead architects of the original research proposal. In addition the research team is indebted to our international expert advisors—Professor George Galster (Wayne State University, Michigan) and Professor Peter Williams (University of Cambridge) for their invaluable advice. DISCLAIMER AHURI Limited is an independent, non-political body which has supported this project as part of its program of research into housing and urban development, which it hopes will be of value to policy-makers, researchers, industry and communities. The opinions in this publication reflect the views of the authors and do not necessarily reflect those of AHURI Limited, its Board or its funding organisations. No responsibility is accepted by AHURI Limited or its Board or its funders for the accuracy or omission of any statement, opinion, advice or information in this publication. AHURI FINAL REPORT SERIES AHURI Final Reports is a refereed series presenting the results of original research to a diverse readership of policy-makers, researchers and practitioners. PEER REVIEW STATEMENT An objective assessment of all reports published in the AHURI Final Report Series by carefully selected experts in the field ensures that material of the highest quality is published. The AHURI Final Report Series employs a double-blind peer review of the full Final Report where anonymity is strictly observed between authors and referees.

ii

CONTENTS LIST OF TABLES ..................................................................................................... VI LIST OF FIGURES ...................................................................................................VII ACRONYMS ............................................................................................................VIII EXECUTIVE SUMMARY ............................................................................................ 1 1 INTRODUCTION ............................................................................................... 5 1.1

Overview ............................................................................................................ 5

1.2

Research context ............................................................................................... 6 1.2.1 Housing, space and place ......................................................................... 6 1.2.2 Disadvantage and housing tenure............................................................. 6 1.2.3 Locational disadvantage and social exclusion ........................................... 7 1.2.4 Concentrations defined by place but shaped across different spatial scales 8 1.2.5 Neighbourhood effects and social mix....................................................... 8 1.2.6 Linkage with official policy priorities .......................................................... 9 1.2.7 Institutional setting .................................................................................... 9

1.3

Research methods and associated research outputs ....................................... 10 1.3.1 Review of Australian and international research and policy literature ...... 12 1.3.2 Statistical analysis and mapping of socio-economic disadvantage .......... 12 1.3.3 Analysis of disadvantaged area housing markets.................................... 13 1.3.4 Exploratory analysis of spatial impacts of housing expenditure ............... 13 1.3.5 In-depth local studies of selected disadvantaged places ......................... 13 1.3.6 Survey of disadvantaged locality residents.............................................. 14

1.4

Report structure ............................................................................................... 14

2

CONCEPTUALISING SPATIAL DISADVANTAGE ......................................... 15

2.1 2.2

Introduction ...................................................................................................... 15 Disadvantage ................................................................................................... 15

2.3

Spatial disadvantage ........................................................................................ 17

2.4

Urban space and place .................................................................................... 18 2.4.1 Space and place ..................................................................................... 18 2.4.2 Configuring urban space and place ......................................................... 19 2.4.3 Flows ...................................................................................................... 20

2.5

The importance of scale ................................................................................... 21 2.5.1 Dysfunctional disadvantage experienced at a small scale ....................... 21 2.5.2 Locational disadvantage in a city-wide or city region context .................. 22 2.5.3 Concentration of socio-economic disadvantage: a problem of scale ....... 22

2.5.4 Some issues ........................................................................................... 23 2.6 Drivers of spatial disadvantage ........................................................................ 24 2.6.1 Urban governance .................................................................................. 24 2.6.2 Labour markets ....................................................................................... 25 2.6.3 Housing market restructuring .................................................................. 26 2.7 Summary ......................................................................................................... 27

iii

3

THE CHANGING GEOGRAPHY OF DISADVANTAGE: INDIVIDUAL AND URBAN SYSTEM IMPACTS ........................................................................... 29

3.1

Chapter overview and sources ......................................................................... 29

3.2

Analysing the geography of socio-economic disadvantage in Sydney, Melbourne and Brisbane .................................................................................................... 29

3.3

The changing spatial pattern of socio-economic disadvantage......................... 31 3.3.1 Overview ................................................................................................. 31 3.3.2 The growing spatial concentration of socio-economic disadvantage ....... 31 3.3.3 The dynamics of socio-spatial change .................................................... 33 3.3.4 The ongoing suburbanising dynamic: recent trends ................................ 33 3.3.5 The ongoing suburbanising dynamic: drivers of change .......................... 34 3.3.6 Housing market processes and spatial disadvantage: evidence from this study ....................................................................................................... 36

3.4

Places that disadvantage people ..................................................................... 38 3.4.1 The ‘neighbourhood effects’ debate ........................................................ 38 3.4.2 Impacts of spatially concentrated disadvantage: evidence from this study 40

3.5

Impacts of the observed geography of concentrated disadvantage .................. 44

3.6 4

Chapter summary ............................................................................................ 45 ADDRESSING CONCENTRATIONS OF DISADVANTAGE: POLICY RESPONSES .................................................................................................. 47

4.1 4.2

Chapter overview ............................................................................................. 47 Conceptualising policy responses to spatial disadvantage ............................... 48

4.3

The anatomy of spatial disadvantage in Australia: comparing six localities ...... 50

4.4

Policy interventions for addressing concentrated disadvantage ....................... 54 4.4.1 Place-focused/person impact policies (Policy type 2) .............................. 54 4.4.2 Person-focused/place impact policies (Policy type 3) .............................. 56 4.4.3 Integrated area-based policies (Policy type 5)......................................... 56 4.4.4 Place-focused/place impact policies (Policy type 1) ................................ 57

4.5

Assessing the effectiveness of place-oriented policies: best for people and best for place? A critical review ............................................................................... 58 4.5.1 The contested evidence for neighbourhood effects ................................. 59 4.5.2 The gentrifying effects of regional growth strategies and place improvement ........................................................................................... 59 4.5.3 People-targeted policies and the unintended outcome of selective migration ................................................................................................. 60 4.5.4 Tacking the structural causes of poverty and disadvantage .................... 60

4.6

A defence of spatially-targeted approaches ..................................................... 62

4.7

Principles for addressing concentrated disadvantage: the importance of remaining ‘place conscious’ ............................................................................. 63

4.8 5

Chapter summary ............................................................................................ 65 CONCLUSIONS .............................................................................................. 66

5.1

Key findings from the research ......................................................................... 66 5.1.1 Urban restructuring and spatially concentrated socio-economic disadvantage .......................................................................................... 66 iv

5.1.2 The role of housing markets in urban spatial restructuring ...................... 66 5.1.3 The consequences of living in a disadvantaged place ............................. 68 5.1.4 The spatial distribution of housing-related expenditure ........................... 68 5.1.5 Spatially concentrated disadvantage: A cause for policy concern? ......... 68 5.1.6 Policy responses to spatial concentrations of disadvantage .................... 69 5.2

Relevance of research findings beyond Australia ............................................. 71

5.3

Reflections on the research methodology ........................................................ 71 5.3.1 Selection of analytical scale .................................................................... 71 5.3.2 Utility of SEIFA in measuring income deprivation .................................... 72 5.3.3 Alternative approaches to typologising disadvantaged places ................. 72 5.3.4 Housing market analysis ......................................................................... 72 5.3.5 The importance of the ABS Census of Population and Housing .............. 72

5.4

Unanswered questions..................................................................................... 73

REFERENCES ......................................................................................................... 74 APPENDICES ........................................................................................................... 85 Appendix 1: Maps of disadvantaged suburb locations and types............................... 85 Appendix 2: Case study areas, selected demographic, employment and housing characteristics .................................................................................................. 92 Appendix 3: House prices and rents analysis—updated to 2014 for Sydney ............. 95

v

LIST OF TABLES Table 1: Sydney, Melbourne and Brisbane: population size and recent change ........ 10 Table 2: Suite of research reports generated by AHURI Project 70704 ‘Addressing Concentrations of Disadvantage’........................................................................ 11 Table 3: Government ‘expenditure’ programs included in spatial analysis ................. 13 Table 4: Calibrating the 2006 geography of disadvantage in Sydney, Melbourne and Brisbane............................................................................................................. 30 Table 5: Disadvantaged suburbs typology in socio-economic and housing market terms .................................................................................................................. 31 Table 6: Comparing 2006 and 2011 spatial concentration of disadvantage in Sydney, Melbourne and Brisbane .................................................................................... 32 Table 7: Average distance from CBD to disadvantaged suburbs (kilometre) ............. 34 Table 8: Neighbourhood effects typology .................................................................. 39 Table 9: Respondents' views on their locality by housing tenure ............................... 41 Table 10: Disadvantaged area accessibility to services and employment: residents’ perceptions ........................................................................................................ 43 Table 11: Summary of the nature of place- and people-based disadvantage in different localities ............................................................................................... 50 Table 12: Dimensions of place disadvantage by suburb type—as exemplified by case study suburbs..................................................................................................... 53 Table 13: Summary of policy interventions across the study localities ....................... 54

Table A1: Demographic characteristics, percentage difference from respective Greater Metropolitan Area, 2011 ........................................................................ 92 Table A2: Employment and education characteristics, percentage difference from respective Greater Metropolitan Area, 2011 ....................................................... 93 Table A3: 2011 Housing costs in case study suburbs ($) .......................................... 93 Table A4: Percentage of housing costs in disadvantaged suburbs relative to city means. Median entry rents by dwelling type (2001–11)...................................... 94 Table A5: Sales prices—detached houses ................................................................ 95 Table A6: Sales prices—other dwellings ................................................................... 95 Table A7: Entry rents: three–bedroom detached houses ........................................... 96 Table A8: Entry rents—one to two-bedroom ‘other dwellings’ ................................... 96

vi

LIST OF FIGURES Figure 1: Policy objectives and targeting in relation to person and place, with examples ........................................................................................................... 49

Figure A1: Disadvantaged suburbs in Sydney: lowest and second lowest decile threshold levels .................................................................................................. 85 Figure A2: Disadvantaged suburbs in Melbourne: lowest and second lowest decile threshold levels .................................................................................................. 86 Figure A3: Disadvantaged suburbs in Brisbane: lowest and second lowest decile threshold levels .................................................................................................. 87 Figure A4: Disadvantaged suburbs in Sydney differentiated according to socioeconomic variables (2001 and 2011) ................................................................. 88 Figure A5: Disadvantaged suburbs in Melbourne differentiated according to socioeconomic variables (2001 and 2011) ................................................................. 89 Figure A6: Disadvantaged suburbs in Brisbane differentiated according to socioeconomic variables (2001 and 2011) ................................................................. 90 Figure A7: Sydney disadvantaged suburbs residents survey fieldwork locations ....... 91

vii

ACRONYMS ABIs

Area based Iniatives

ABS

Australian Bureau of Statistics

AHURI

Australian Housing and Urban Research Institute Limited

AURIN

Australian Urban Research Infrastructure Network

BD

Central Business District

CDs

Collection Districts

COAG

Council of Australian Governments

CPPs

Critical Perspective Papers

CT

Census Tract

GMAs

Greater Metropolitan Areas

IRSD

Index of Relative Socio-economic Disadvantage

LGAs

Local Government Areas

NSW

New South Wales

QLD

Queensland

SEIFA

Socio-Economic Indexes for Areas

VIC

Victoria

viii

EXECUTIVE SUMMARY Key findings  Socio-economically disadvantaged populations of Australia’s major cities are substantially

clustered into suburbs now predominantly located in middle and outer metropolitan areas.  Disadvantaged places in Sydney, Melbourne and Brisbane encompass significant diversity,

with these suburbs separable into four socio-economically distinct types. These discrete archetypes are also distinctive in terms of housing market structures and characteristics.  The increased clustering of disadvantaged neighbourhoods evident in the 2006–11 period

extends a tendency ongoing since the 1980s and compounds other evidence of growing economic marginalisation of lower status suburbs in Australia’s major cities.  While first highlighted in Australia, the suburbanisation of socially disadvantaged

populations is an international trend—increasingly also recognised in Europe and North America. However, by comparison with the UK and the US, it is in Australia’s major cities where the process has impacted most decisively.  As the least mobile component of the workforce has been increasingly consigned to areas

remote from the inner city epicentre of the recently burgeoning ‘knowledge economy’, the economic exclusion of lower income urban Australians is likely to be compounded.  Alongside its acknowledged negative social and economic impacts for residents directly

impacted, there is a dawning recognition that growing spatial polarisation of our major cities impairs overall urban productivity, thus imposing costs on all.  The evolving geography of disadvantage in urban Australia largely reflects the outcome of

polarising market forces, virtually unmitigated by countervailing policy intervention. However, as quantified by the research, the spatial distribution of housing-related government expenditure is profoundly regressive—proportionately far greater ‘housing subsidies’ flow to advantaged postcodes than to disadvantaged postcodes.  From a metropolitan planning perspective, the research highlights the need for:

1. City-scale strategies to encourage growth nodes (including employment, public facilities and services, cultural institutions) in outer suburban locations. 2. Effective measures to expand and protect affordable housing in existing well-located areas and localities proximate to new decentralised employment nodes where, in the absence of intervention to counter market processes, housing costs will be bid up to the detriment of low-income groups.

Background Across the developed world, ongoing growth in intra-national income disparities has been recently paralleled by heightened concerns at the harmful effects of such trends (Wilkinson & Pickett 2009; Picketty 2014; IMF 2014). Spatial disadvantage is the socially polarising process through which growing economic inequality is expressed and embodied in the geography of human settlements. Contemporary anxieties about spatially unequal cities stem partly from the belief that an individual’s place of residence affects their life chances (Vinson 2009; Galster 2012). Compounding such ‘social’ concerns there is a growing recognition that growing geographical segregation also impairs city productivity, thereby imposing costs on all (Kelly & Mares 2013). This is the final publication in a suite of reports drawing on a research program about spatial disadvantage in urban Australia, the experience of living in a ‘disadvantaged area’ and the role of housing markets and housing policies in contributing to, and mitigating, spatial disadvantage.

1

Focusing on Australia’s three largest capital cities, Sydney, Melbourne and Brisbane, overarching issues covered by the research were:  How concentrated disadvantage is conceptualised and how housing and urban systems

contribute to evolving spatial patterns.  The impacts of spatially-concentrated disadvantage for directly affected residents and for

urban systems.  The policy implications stemming from the spatially concentrated socio-economic

disadvantage and its changing geography.

Conceptualising spatial disadvantage Establishing a clear understanding of spatial disadvantage is vital in informing Australia’s urban policy debate.

Interpreting ‘disadvantage’ Disadvantage 1 is an umbrella notion embracing a series of concepts (including poverty, deprivation, social exclusion and social capital), which describe different aspects of distributive and/or social inequality, referring respectively to the distribution of economic resources and other disparities between people or population groups. Spatial disadvantage 2 is likewise an umbrella term incorporating three concepts: locational disadvantage, places with social problems, and concentrated disadvantage, the latter referring to places accommodating a disproportionate number of socio-economically disadvantaged people. Spatial disadvantage, like disadvantage generally, results from both distributive and social inequality, and is relative and multi-dimensional. While relating to people with low incomes, it also includes a relative lack of access to resources and opportunities. It is best seen as a dynamic process rather than an end state. In understanding spatial disadvantage, it is important to unpack the relationship between the restructuring of urban space and the making of particular places. The restructuring of urban space through market transactions and public policies shapes human settlements, but places are also made by people as they interact with others, select the locales for their everyday activities, and ascribe meanings to place. It is also important to consider cities as ‘flows’ as well as ‘stocks’; the ways people move about the city and the extent of residential mobility influence the making and remaking of places with relative degrees of advantage and disadvantage.

Scale and diversity Appropriate selection of analytical scale is critical in identifying and investigating spatial disadvantage. Places with social problems are best seen at a fine spatial scale—perhaps the ‘neighbourhood’. Locational disadvantage is best viewed at a larger urban or city region scale although it may have highly localised impacts. Detecting spatially concentrated socio-economic disadvantage is more problematic since choice of scale influences whether such concentrations are, in fact, identified. We used suburb as our preferred spatial framework, since the neighbourhood is too small and imprecise a unit for this purpose. Spatial disadvantage takes different forms and not all disadvantaged places are the same. They play a variety of roles within their broader spatial contexts. These different drivers and contexts necessarily flow through to different policy responses.

1

We consider this best described as either disadvantage, or socio-economic disadvantage, to include incorporate dimensions of distributional and social inequality; it applies to people rather than places. 2 We consider this best described as spatial disadvantage—including both economic and social aspects—rather than socio-spatial disadvantage which downplays economic drivers and impacts.

2

The changing geography of socio-economic disadvantage: individual and urban system impacts Perhaps the most telling manifestation of growing spatial polarisation in urban Australia is the evidence that, over the past 40 years, unemployment has become markedly more concentrated in lower status suburbs. There are similarities between the urban spatial dynamics underlying such processes as seen in Australia and as recognised in other comparable countries (e.g. in the UK and the US). However, our distinct social and urban heritage means that there are important contrasts with these comparator countries. While continuing to grow, the spatial concentration of disadvantage remains less intense in Australia. Thus, Australian spatial disadvantage may be relatively ‘shallow’ yet the greater fluidity of the nation’s urban systems is evident in that geographically focused disadvantage has moved more decisively from inner urban localities to middle and outer ring suburbs than in comparator British or American cities. Detailed studies of disadvantaged suburbs found high levels of economic exclusion featuring jobless rates (especially for young people) well above city-wide norms. As demonstrated by our own survey evidence, residents of middle and outer suburban disadvantaged places are inclined to view their location as problematic mainly in terms of access to employment. Such perceptions are symptomatic of the escalating spatial mismatch between lower income populations—the least mobile workforce segment—and the predominantly inner city employment growth centres of the emerging ‘knowledge economy’. This is embodied in the observed suburban economic participation deficit, especially affecting women. For individuals, the result is unfulfilled potential; for society, the impact is suppressed city productivity. Thus, the intensifying exclusion of lower income populations from inner cities is compounding settlement patterns detrimental to city productivity as well as to individual welfare for those directly concerned. Addressing concentrations of disadvantage: policy responses By analysing policy interventions designed to address spatial disadvantage in selected study neighbourhoods, the research considered whether, and how, interventions of this type can be both ‘best for people’ and ‘best for place’. Integrated area-based policies have become subject to critique in recent years for failing to fully address the core features of poverty and disadvantage—limited employment opportunities, poor educational achievement and housing unaffordability. They are also considered problematic because they create inequality between disadvantaged areas targeted with local area renewal schemes and those omitted, and because they fail to support disadvantaged populations resident in non-disadvantaged places. In Australia, where area-based policies have been largely absent, except as applied to deteriorating public housing estates, the ability to combat the broader economic and political drivers of disadvantage has been even more limited. In any case, area-based interventions alone cannot address wider systemic problems. There is a ‘scalar mismatch’ (Rae 2011) when local policy solutions are used to address systemic causes and consequences of disadvantage that have their genesis in national, regional or urban policies and processes. Such policies inevitably fail to meet their ‘best for people; best for place’ ambitions. Governments should therefore consider whether targeted investments into neighbourhoods might be best channelled through a-spatial policies of income, education and housing support. It is also important to be alert to the spatial impact of housing and other policies. According to our analysis of expenditure flows in Melbourne households in the top quartile, most advantaged postcodes received, on average, $4600 in direct and indirect government housingrelated benefits in 2011–12, while households in the bottom quartile postcodes received, on average, $2800. While exploratory in nature, and confined to one city, this emphasises serious questions about the net spatial impacts of ’housing subsidies’, where the indirect benefits

3

largely accruing to already advantaged households through the tax system far outweigh the direct impacts of public housing and other ‘pro-poor’ programs. Housing markets ‘sift and sort’ where people live. While there is little evidence that the concentration of socio-economically disadvantaged people per se has an independent effect on people’s lives, place still matters particularly where concentration of disadvantage overlaps with experience of social problems and locational disadvantage. Policies need to be sensitive to the particular forms that disadvantage takes in specific local settings while also recognising the broader structures and processes that affect all such areas in similar ways. Policy-makers should also recognise, as highlighted by our research, that ‘disadvantaged suburbs’ can also be diverse and vibrant communities.

Conclusions The research findings suggest that, by comparison with some other developed countries, especially the USA, the depth of spatial disadvantage found in Australia’s major cities remains moderate rather than extreme. Thus, within Australia’s urban ‘poverty concentration areas’ the incidence of social and economic deprivation is well above city-wide or national norms yet far from universal. While the Australian tendency towards ‘suburbanisation of disadvantage’ is in keeping with international trends, it has impacted more decisively here than elsewhere. This is of concern in view of the global rise of the ‘knowledge economy’, burgeoning in well-connected inner urban areas. This is a particular issue in Australia’s geographically large and low density cities, with their largely mono-centric structures. While a range of policy domains are potentially relevant to remediating locational disadvantage (e.g. transport policy and metropolitan planning), and places with social problems (e.g. justice and human services), the concentration of disadvantage poses particular challenges for policymakers. Public, as well as private, investment in inner cities has contributed to the steepening house price/rent gradient between these localities and the outer suburbs, as seen since the 1980s. The ‘hidden hand’ of public policy arguably tends to compound market processes here, but there are ‘moderating’ interventions that could be considered—for example through mandating affordable housing provision in accessible locations via inclusionary zoning, through targeting transport and other productive infrastructure investment to promote economic decentralisation from CBDs to secondary growth nodes, or through greater investment in cultural infrastructure in non-CBD locations.

4

1

INTRODUCTION

1.1

Overview

Across the developed world, ongoing growth in intra-national income disparities has been recently paralleled by heightened concerns at the harmful effects of such trends (Wilkinson & Pickett 2009; Picketty 2014; IMF 2014). Spatial disadvantage is the socially polarising process through which growing economic inequality is expressed and embodied in the geography of human settlements. Contemporary anxieties about spatially unequal cities stem partly from the belief that an individual’s place of residence affects their life chances (Vinson 2009; Galster 2012). Compounding such ‘social’ concerns there is a growing recognition that growing geographical segregation also impairs city productivity, thereby imposing costs on all (Kelly & Mares 2013). This publication is the final output in a suite of reports drawing on the ‘Addressing Concentrations of Disadvantage’ research program that commenced in 2011 and was completed in 2014. This major study aimed to investigate the geography of socio-economic disadvantage in urban Australia, the experience of living in a ‘disadvantaged area’ and the ways that housing policies and processes are implicated in the changing patterns observed. The overarching issues covered by the research were defined as follows:  How concentrations of spatial disadvantage have been conceptualised and how this relates

to our broader understanding of the operation and impacts of housing and urban systems.  The impacts of spatial disadvantage, and the importance of housing and place in mediating

the incidence and experience of residents of disadvantaged areas.  How policy, practitioners and communities can respond to spatial disadvantage in ‘best for

people, best for place’ terms. The study was designed to play into Australian and international debates on social exclusion and place-based disadvantage and, in particular, to stimulate critical engagement with the spatiality of drivers, impacts and outcomes affecting exclusion and how they are mediated by place. Underlying this was the aspiration to develop a more considered understanding of the spatial implications of government policies, especially as regards distributional impacts of housing and related programs. Equally, the research was designed to inform policy-making around enhancing urban productivity, an issue that has subsequently assumed greater priority under the post-2013 Australian (Federal) Government.3 Innovative aspects of the research included the integration of small area administrative data and census-derived statistics in identifying concentrations of disadvantage and analysing housing market change in these areas. Also distinctive was the combination of in-depth analysis at a fine spatial scale and the study’s wide-ranging geographical remit, encompassing all three of Australia’s largest capital cities—Sydney, Melbourne and Brisbane. Incorporating a range of quantitative and qualitative techniques, the research was undertaken in a series of stages, generating a series of five freestanding AHURI research reports already available via the project webpage at: http://www.ahuri.edu.au/publications/projects/myrp704. The methodology employed in each of these stages is briefly summarised below (see Section 1.4 and Table 2). This ‘project conclusion’ report revisits our original overarching research questions. In doing so, it integrates relevant evidence derived from each distinct aspect of the fieldwork and analysis. It also draws on a fresh 2014–15 review of relevant Australian and international literature, mainly featuring work published while our study was under way. Additionally, the report reflects 3

Also known as the Commonwealth Government.

5

on the study in terms of methodological lessons and in terms of pointers for the research agenda going forward. The remainder of this chapter is structured as follows. First, we briefly frame the research by highlighting certain key developments and concerns prompting the study and relevant aspects of the evolving policy context for the research. Summarised here are a number of key issues shaping academic and policy interest in spatially concentrated disadvantage and place-based factors which may compound the problems of local residents. Second, we briefly outline— especially for the benefit of non-Australian readers—the institutional setting within which the research was conducted. Third, we outline the research methods used in the study and the remit of the individual reports published to date. Finally, we outline the structure of the remainder of this report.

1.2

Research context

1.2.1 Housing, space and place It has been long recognised that such economic processes, are reflected in the evolving socialspatial patterning of the urban landscape (Badcock 1997; O’Connor et al. 2001; Gleeson 2005). The increased reach of globalising forces, economic restructuring and the transformed nature of labour market dynamics act to sift and sort opportunities and constraints for households, shaping and reinforcing spatial outcomes. This connects with an enduring academic interest in what has been termed ‘socio-spatial polarisation’,4 highlighting the social restructuring of urban space (Badcock 1984; Forster 1995; Vinson 2004; Randolph 2006; Baum & Gleeson 2010). While social inequality is an evergreen research topic, interest in the role of housing, ‘place’ and neighbourhood has come increasingly to the fore, both in terms of understanding the genesis of, and solutions to, spatial disadvantage. A ‘spatial turn’ in social science has focused attention on the impacts of place on social and economic prospects of individuals (Darcy & Gwyther 2012). Housing as a site and focus of welfare transitions (Smith 2009) has had ‘a profound … effect on the spatial distribution of outcomes mediated principally through housing and economic policy’ (Lee 2010, p.185). Equally, recognition that there is an important ‘place dimension’ to the experience of poverty remains far from universal in Australia’s policy-making community. For example, a major recently-published study on ‘deep and persistent disadvantage’ contained only the briefest reference to spatial considerations (Productivity Commission 2013). Housing markets, public policies and the ways in which people consider, and make decisions about, their housing interact to locate people in particular places. Places have multiple dimensions: they are physical entities as well as sites of economic transactions, social relations and cultural significance (Hulse et al. 2011, p.18). Housing policy settings have been complicit through further complicating the ‘basket of goods’ nature of the home to accommodate a raft of consumption, investment and wealth concerns (Burke & Hulse 2010). Therefore, especially given the pivotal role of housing markets in driving socio-spatial sifting (Randolph & Holloway 2005b), better comprehension of socio-spatial polarisation calls for improved understanding of the function of housing and its relationship across space.

1.2.2 Disadvantage and housing tenure In many developed countries, the effective withdrawal of funding for new social housing over recent decades has reinforced socio-tenurial polarisation (Hamnett & Randolph 1986); consequentially shrinking social rental systems have become further dominated by highly disadvantaged populations. As manifested in Australia, this has resulted in a problematic public housing sector reality visually identifiable to policy-makers: asset holdings concentrated in 4

As explained in Chapter 2, we prefer the term ‘spatial disadvantage’.

6

large estates result in associations between poverty and place becoming a political and a policy issue (Badcock 1997; Winter & Stone 1998; Peel 2003). Housing policy is therefore caught in a contradictory bind: ethically prioritising access to sub-market accommodation for those in greatest need, yet seeking to reduce concentrations of poverty. Associations between public housing and disadvantage have historically formed the overwhelming focus for Australian policy interest in this area. This is highlighted by the aspiration of ‘creating mixed communities that promote social and economic opportunities by reducing concentrations of disadvantage that exist in some social housing estates’, included as a key aim in the 2009 National Affordable Housing Agreement (COAG 2009, p.7). Clearly, this echoes the ‘poverty dispersal’ ethic which has underpinned the mass demolition of US public housing estates—especially under the HOPE VI program of the 1990s and 2000s (Goetz 2013). As further discussed in Section 3.4.1, the merits and intellectual underpinning of this policy have been hotly debated (Darcy & Gwyther 2012; Slater 2013; Schwartz 2015). Critics have also questioned the relevance of this experience to Australia (Arthurson 2013), and it should be acknowledged that state and territory governments have, as yet, initiated relatively little wide-scale clearance of public housing estates and managed the scattering of former residents. At the time of writing, large-scale renewal projects incorporating ‘public housing de-concentration’ remain unusual. Even well-known large public housing redevelopment projects such as those at Bonnyrigg (NSW) and Kensington (Victoria) aimed to ‘de-concentrate’ social renters only to the extent of re-providing (on-site) social rental dwellings in numbers similar to their original representation, within the context of new ‘densified’ layouts where additional privately owned dwellings ‘dilute the mix’. Moreover, the overarching significance of racial politics in US urban policy is unparalleled in Australia. However, given the configuration of Australia’s housing system (in which 95% of housing is in the private market), much of the country’s low-income population resides in the private housing market—both among private renters and outright home owners—rather than in public housing. Thus, shifting geographies of disadvantage in our major cities have been substantially associated with the evolving pattern of private rental housing provision. Of critical importance here—at least in Sydney—has been a long-term tendency towards the increased spatial concentration of low (private) rent dwellings, effectively reinforcing the socio-spatial polarisation of the city (Yates & Wood 2005). Undoubtedly linked with this, in Sydney and other large cities, the locations of socio-economic disadvantage have shifted decisively outwards in over the last 20–30 years (Randolph & Holloway 2005b). While once largely an inner-city issue, the spatial concentration of low-income groups is now substantially a problem of middle and outer suburban areas (Hulse & Saugeres 2008; Burke & Stone 2014). Housing, planning and urban policy frameworks in Australia have, nevertheless, remained essentially silent on places where such concentrations are cross-tenure or predominantly composed of private sector housing. Here, public housing departments and public housingfocused governments recognise no remit to act. In the absence of a ‘whole-of-housing-system’ perspective5 there has been little recognition of the importance of space and place, and the role and function of localities in their broader housing and labour market contexts.

1.2.3 Locational disadvantage and social exclusion This study builds on an established tradition in Australian urban studies in which concerns about ‘locational disadvantage’ have focused on attributes of localities negatively impacting on residents’ life chances (Maher et al. 1992; Badcock 1994). These perspectives were linked with David Harvey’s view that '… place specific urban resources such as employment opportunities, 5

This is a reference to the phraseology of the Australian Government’s 2008 National Affordable Housing Agreement (NAHA). In its aspiration to improve housing affordability through a ‘whole of housing system’ approach, Government was signaling a recognition that policy needed to stretch beyond its traditional ‘public housing’ preoccupation.

7

changes in property value and the availability and price of other resources are ‘fringe benefits’ that contribute to the real income of households’ (Harvey 1973, p.8). In one critical view, indeed, the idea that ‘where you live affects your life chances’ commands such support that it has led to ‘analytic hegemony’ in urban studies (Slater 2013). Locational disadvantage is generally conceived as attributable to residential remoteness from employment and services—for example dependent on whether the locality is serviced by CBDconnected road or rail links. However, locational disadvantage is a multi-scalar concept. Thus, particularly in relation to car-less low-income households, there are gradations even within suburbs. Such residents living beyond walking distance of a station or good bus route will experience another level of locational disadvantage. Further, while locational disadvantage tends to be thought about mainly within the context of ‘remoteness from employment and services’, other forms of the problem could include a suburb’s proximity to excess noise (e.g. road traffic or airport flightpath) or other forms of pollution, or susceptibility to other environmental hazards such as flooding. Recent policy thinking has been strongly influenced by the concept of social exclusion. Argued as a broader notion than ‘income poverty’, this emphasises individuals’ connectedness to social networks and labour markets (Arthurson & Jacobs 2003). Linked with this is the concept of spatial exclusion and the idea of ‘exclusionary places’ (Sibley 1996)—localities described as developing their own ‘pathologies’. Indeed, it is argued that ‘the metaphor of ‘exclusion’ has contributed to the construction of disadvantage as a largely spatial phenomenon’ (Darcy & Gwyther 2012, p.250).

1.2.4 Concentrations defined by place but shaped across different spatial scales In focusing on concentrations of disadvantage, rather than disadvantage per se, questions of scale, boundaries and composition are central. In countries such as the UK, but also to some extent in Australia, the ‘neighbourhood’ has been a traditional focus of place-based policy intervention. This level of geography ‘represents a scale at which many government services and provisions are made … The neighbourhood is a scale at which people can be persuaded to get involved and feel a sense of belonging’ (Manley et al. 2013, p.3). In analysing problems at this scale there is a need to understand local housing markets, economic function and population mobility in spatial terms to develop appropriate policy responses.

1.2.5 Neighbourhood effects and social mix A focus on place, and concentrations in place, has stimulated extensive academic and policymaker interest in whether geography, proximity and association compound or mitigate disadvantage. Studies have considered compositional factors, whether ‘thresholds’ can be identified beyond which point neighbourhood effects act to compound negative externalities and impact on wellbeing, opportunities and house prices (Buck 2001; McCulloch 2001). Whether living in a poor area significantly compounds the problems of poor individuals has been widely researched, particularly in the US, with great weight attaching to evaluations of ‘assisted mobility’ programs designed to disperse low-income populations. The topic, however, remains vigorously contested. This derives partly from the way that policy-makers (especially in the US) are understood to have cited ‘neighbourhood effects’ in justifying the so-called ‘deconcentration’ of public housing estates—a policy which some believe amounts to ‘state sponsored gentrification’ to the detriment of the dispersed communities themselves. Other than as regards stigmatisation of social housing as a tenure form (e.g. Jacobs et al. 2011), there has been little empirical evidence of these effects in an Australian context. Closely associated with debates on ‘neighbourhood effects’ is the academic and policy-maker contestation around ‘social mix’ as a key aspiration of land use planning, and around the elision of social mix and tenure mix in official discourse. The absence of strong underpinning evidence on beneficial impacts of such heterogeneity has been highlighted by many (e.g. 8

Arthurson 2013; Shaw & Hagemans 2015). However, this particular stream of literature is less directly relevant to the current research, since our primary focus is areas containing concentrated disadvantage rather than those where ‘mixed communities’ ideology has been put into practice to any significant extent.

1.2.6 Linkage with official policy priorities All Australian (Federal) governments in the 2000s have placed faith in enhanced economic participation to improve individual and family financial capacity, as well as to boost economic productivity in the face of population ageing (e.g. Australian Government 2002, 2010). The former Labor Australian Government framed this in terms of a social inclusion strategy (2009– 13) which set important markers in terms of better cross-government coordination and identification of place-based disadvantage. Implicit here was a need for rigorous engagement and understanding of the spatiality of exclusion drivers, impacts and outcomes and how these are mediated by place. Under the post-2013 Australian Government, however, the policy primacy of social inclusion has been eclipsed by aspirations to reduce ‘welfare dependency’ and ‘maximise urban productivity’ (Australian Government 2011; Kelly & Mares 2013). From the Federal perspective, therefore, the policy relevance of spatially concentrated disadvantage is now seen largely through this prism.

1.2.7 Institutional setting Especially for the benefit of overseas readers, we should briefly note the institutional context for urban and housing policy in Australia. Under the country’s federal governance framework 6 the constitutional responsibility for policy formulation and implementation in these specific realms rests primarily with the states and territories. From time to time, national government has acted more assertively in these areas (e.g. in the mid-late 1940s, the early-mid 1970s, the early 1990s and between 2007 and 2013. More customarily the direct role of the Australian (Federal) Government in housing and urban policy has been through contributing to funding the states/territories for relevant activities and exerting associated leverage in seeking to influence their implementation. Crucially framing these agendas, however, are the Australian Government positions and policy initiatives on tax, social security and migration rules for which Canberra retains virtually all responsibility. Other than through conditions attached to funding agreements, the Commonwealth Government has little direct power over urban and housing policies. Most legislation governing these matters is state-specific and relevant funding decisions lie with ministers in state and territory administrations. It is at this level of government that public housing is operated and land use planning policy determined. While responsible under delegation of state powers for land use planning and implementation, local government plays a relatively minor role in urban and housing policy. Arguably, this governance tier is hampered not only through its limited powers and resource base, but also due to its fragmented and disparate nature. Across Greater Sydney and Greater Melbourne, for example, there are currently 69 municipalities—many operating at a very small scale. Nevertheless, some local authorities actively promote community development and participate in multi-agency partnerships which contribute to the social and/or economic renewal of disadvantaged areas.

6

There are six states and two territories (the Northern Territory and the Australian Capital Territory, the latter including Canberra, the nation’s capital). Local government has no specific powers under the Australian Constitution and implements programs though delegated state/territory powers. Local government areas across Australia (with the notable exception of Brisbane) cover relatively small areas.

9

1.3

Research methods and associated research outputs

The study focused on Australia’s three largest cities: Sydney, Melbourne and Brisbane. As shown in Table 1 below, these metropolitan areas accounted for almost half the national population in 2013. This is part of a wider norm in which two-thirds of all Australians live in state capital cities. Importantly, however, Australia’s geography is such that these exist as largely self-contained housing markets with relatively little inter-city mobility. One significant implication is that lower income households pressured by unaffordable housing costs have tended to move outwards towards (or slightly beyond) city margins rather than shifting to an entirely new urban area. For our purposes, the three cities were defined according to their ABS ‘statistical division’ boundaries which relate to cities as economic units that extend beyond the edge of current built-up areas. Consequently, the areas concerned extended to some 60 kilometres from respective Central Business Districts (CBDs). Table 1: Sydney, Melbourne and Brisbane: population size and recent change Population numbers (million)

% of national total

% change 2003–13

2003

2013

2003

2013

Sydney

3.8

4.4

19.3

18.9

14.6

Melbourne

3.5

4.2

17.6

18.1

20.7

Brisbane

1.7

2.1

8.7

9.3

25.5

Rest of Australia

10.7

12.4

54.4

53.8

15.9

Australia—total

19.7

23.1

100.0

100.0

17.3

Source: Australian Bureau of Statistics, cat no. 3218.0

The study involved an integrated set of research methods. The six main elements, together with their associated reports—as separately published by AHURI—were as listed in Table 2 below. The remainder of this section summarises the methodology and data sources involved in each of these study components. Fuller accounts of how the research was undertaken can be found in each of the listed reports themselves.

10

Table 2: Suite of research reports generated by AHURI Project 70704 ‘Addressing Concentrations of Disadvantage’ Published report details

Research element

Research type

Key questions/aims addressed

Geographic scope

Pawson, H, Davison, G & Wiesel, I 2012, Addressing concentrations of disadvantage: policy, practice and literature review; Final Report no. 190, AHURI, Melbourne.

Review of Australian and international research and policy literature on concentrated disadvantage

Systematic review (qualitative)

What are the evolving academic and policy perspectives on: (a) the processes that lead to concentrations of disadvantage; (b) the processes that contribute to negative outcomes for people living in disadvantaged areas, and (c) urban policy responses to concentrations of disadvantage.

Australian and international

Hulse, K, Pawson, H, Reynolds, M & Herath, S 2014, Disadvantaged places in urban Australia: analysing socioeconomic diversity and housing market performance; Final Report no. 225, AHURI, Melbourne.

Statistical analysis and mapping of socio-economic disadvantage

Quantitative— secondary data analysis

(a) What is the spatial pattern of disadvantage across Australia’s major capital cities? (b) How can we capture the heterogeneity of disadvantaged places in Australia? (c) To what extent are there similarities in the heterogeneity of disadvantaged places and spatial patterns of disadvantage across the three cities?

Sydney, Melbourne and Brisbane

Analysis of disadvantaged area housing markets

Quantitative— secondary data analysis

(a) What are the housing market structures, conditions and dynamics of disadvantaged places? (b) What have been the recent housing market trajectories of disadvantaged places? (c) Can housing market conditions and trajectories be ‘mapped onto’ types of disadvantaged area, distinguished from one another in terms of socio-economic factors?

Sydney, Melbourne and Brisbane

Groenhardt, L 2014, Understanding the spatial impacts of direct and indirect government housing expenditure, Final Report no. 234, AHURI, Melbourne.

Exploratory analysis of spatial impacts of housing expenditure

Quantitative— secondary data analysis

How does the spatial distribution of housing and housing-related government expenditure relate to the spatial distribution of disadvantage?

Melbourne

Cheshire, L, Pawson, H, Easthope, H & Stone, W 2014, Living with place disadvantage: community, practice and policy, Final Report no. 228, AHURI, Melbourne.

In-depth local studies of selected disadvantaged places

Mainly qualitative— indepth interviews and focus groups

(a) To ‘groundtruth’ the spatial typology by piecing together local perceptions of social life within identified suburb types; (b)To complement the residents' survey evidence on residents’ experiences of living in a disadvantaged area; (c)To understand how disadvantage impacts on different social groups (d) To identify local perspectives on the assets and challenges of the locality, and informed views on the area’s socio-economic trajectory; (e) To identify policy and practice responses applied to different areas of disadvantage in the form of people- and placefocused initiatives.

Six disadvantaged suburbs in Sydney, Melbourne and Brisbane: Auburn, Emerton-Mount Druitt (Sydney); Braybrook, Springvale (Melbourne); Logan Central, Russell Island (Brisbane)

Pawson, H & Herath, S 2015, Disadvantaged places in urban Australia: residential mobility, place attachment and social exclusion. Final Report no. 243 Melbourne: AHURI

Survey of disadvantaged locality residents

Household survey of 800 residents

(a) How are disadvantaged places perceived by their residents? (b) How do disadvantaged area housing markets operate and how do housing market processes impact on the spatial concentration of poverty? (c) What is the breadth and depth of social exclusion in disadvantaged places, and how does the incidence of such exclusion vary between different forms of disadvantaged place and across different populations?

Four disadvantaged suburbs in Sydney: Auburn, Emerton (Mount Druitt), The Entrance and Warwick Farm

11

1.3.1 Review of Australian and international research and policy literature As shown in Table 2, AHURI Final Report no.190 was derived from a systematic review of relevant articles, books, book chapters and reports. Around 100 of the identified items originated from Australia, with a similar number from North America, around 50 from the UK and a few (in English) from continental Europe. Selecting from the 250 chosen texts, and giving preference to Australian material, key empirical findings and/or conceptual arguments were distilled into summaries which formed the basis for our wide-ranging literature review as published by AHURI (see Table 2 above). Associated with this work, and to contribute to our thinking on the conceptualisation of ‘disadvantaged places’ as these are manifested in Australia, the following two ‘critical perspectives papers’ (CPPs) were developed to stimulate thinking by the research team:  Burke, T & Hulse, K 2015, Spatial disadvantage: Why is Australia different? Research paper,

AHURI, Melbourne.  Hulse, K & Pinnegar, S 2015, Housing markets and socio-spatial disadvantage: An Australian

perspective; Research paper, AHURI, Melbourne. While our systematic appraisal (see above) and the two CPPs were based on our foundational review completed in 2011–12, the current report benefits from a fresh survey of relevant literature as undertaken in 2014–15.

1.3.2 Statistical analysis and mapping of socio-economic disadvantage As shown in Table 2, AHURI Final Report no.225 drew on a large scale secondary data analysis substantially based on material from the Census of Population and Housing conducted by the Australian Bureau of Statistics (ABS) every five years. The aim was to identify and classify ‘disadvantaged places’ in Sydney, Melbourne and Brisbane. The Census is the only major data set that enables spatially fine-grained disaggregation. While recognising that ‘disadvantaged places’ may be conceptualised in various ways (see Pawson et al. 2012), our analysis adopted a ‘people-centred’ approach under which a ‘disadvantaged place’ was defined as a locality containing a ‘concentration’ of residents subject to socio-economic disadvantage, the line of thinking which has generally underpinned policy-maker concerns in Australia. In terms of geographical scale, the analysis centred on ‘suburbs’, units with a typical population of 4–8000, places which have a socially understood meaning and to which Census and housing market data can be mapped in ways that are not applicable to smaller spatial units. ‘Disadvantaged suburbs’ were classified in relation to the ABS Socio-Economic Indexes for Areas (SEIFA) Index of Relative Disadvantage (IRSD), a statistical measure (in its 2006 version) 7 drawing on 17 variables covering household income, 8 education, employment, occupation, housing and other indicators of disadvantage (ABS 2008). The threshold for designation as a ‘disadvantaged’ locality was a ranking in the lower quintile of the national SEIFA IRSD (hereafter ‘SEIFA’) distribution. In the main, the 177 ‘disadvantaged suburbs’ thus identified were located in middle and outer metropolitan areas of the three cities (see Figures A1–A3 in Appendix 1). Next, to classify the cohort of disadvantaged suburbs, we undertook a cluster analysis encompassing indicators representing social/residential mobility, lifecycle stage/family type and change over time in socio-economic circumstances which were drawn from available and

7

The SEIFA IRSD is one of four indexes calculated by the ABS based on census data. Because of the timing of the research our foundational analysis was necessarily reliant on SEIFA IRSD rankings derived from the 2006 Census— 2011-based rankings had not yet been published. 8 As a measure of disposable income, SEIFA is acknowledged as having important shortcomings. Associated issues are discussed in Section 5.3.2.

12

customised census data for these suburbs. After the exclusion of two outliers, four distinct types of disadvantaged suburbs emerged from this analysis (see Figures A4–A6 in Appendix 1).

1.3.3 Analysis of disadvantaged area housing markets Also encompassed within AHURI Final Report no.225 (see Table 2), this component of the study focused on the cohort of disadvantaged suburbs identified as above, differentiating between the four identified sub-categories. This part of the research drew on original analysis of house sales and private market rents data to analyse housing market performance in the period 2001–11. A central aim here was to calibrate market change in disadvantaged areas, as compared with nondisadvantaged areas as well as investigating difference in housing market performance between the four types of disadvantaged suburbs identified in the cluster analysis. In addition to ABS census data (e.g. on recent mobility), the analysis drew on 2001 and 2011 house sales and rents datasets assembled by the research team for each city from administrative state sources. For house sales, these were Valuer General records9 and for rents at the time of tenancy commencement were from Rental Bond Board10 records. Analysis of these data enabled identification of real change in prices/rents over time as well as the volume of housing market activity.

1.3.4 Exploratory analysis of spatial impacts of housing expenditure As shown in Table 2, AHURI Final Report no.234 analysed the spatial distribution of government spending related to housing. As an exploratory element of the project, this focused solely on Melbourne. Government ‘resource outlays’ covered by the analysis included both direct and indirect expenditures as shown in Table 3. Table 3: Government ‘expenditure’ programs included in spatial analysis Instrument

Direct expenditure

Capital gains tax exemption for owner occupiers First home owner grants and boosts

Indirect expenditure

Responsible tier of government

Yes

Federal

Yes

Negative gearing for rental housing investors

State/territory Yes

Federal

Private rental assistance

Yes

Federal

Public housing provision

Yes

State/territory

Expenditures were calculated to postcode level 11 to reveal the spatial distribution of housing expenditure within Melbourne. The resulting pattern was related to the geography of disadvantage as calibrated via the ABS SEIFA index (see above).

1.3.5 In-depth local studies of selected disadvantaged places AHURI Final Report no.229 (see Table 2) drew on local stakeholder and resident perspectives in a largely qualitative element of the research focused on six disadvantaged suburbs, two each in Sydney, in Melbourne and in Brisbane. Selection of these case study areas, named in Table 2, is fully explained in Final Report no.229. With respect to each case study area, secondary data review and primary fieldwork involved:

9

These were obtained directly in the case of Melbourne and via Australian Property Monitors for Sydney and Brisbane. These are state-based authorities, such as the Residential Tenancies Bond Authority in Victoria. 11 Postcodes form an intermediate tier of geography between suburbs and local government areas (LGAs). Melbourne suburbs averaged some 7000 population in 2006, while postcodes averaged 14 000 and LGAs, 115 000. 10

13

 Desk top review of data, documents and media reports.  In-depth interviews with local stakeholders, typically including the local council, civil society

groups, social housing providers and, in some cases, also encompassing voices such as local police, estate agents, head teachers, etc.  Residents focus group meeting.

Each case study was written up as a freestanding report, all of which have been made available on the AHURI website.

1.3.6 Survey of disadvantaged locality residents As shown in Table 2, AHURI Final Report no.243 covered a residents survey targeted on four disadvantaged suburbs in Sydney. Picking up from the identification and classification of disadvantaged places (see Section 1.4.2 above), this involved face-to-face interviews with 800 residents of four such suburbs—with each chosen locality ‘representing’ one of the four disadvantaged area types. As only Sydney had all four types of disadvantaged suburbs, and partly for logistical reasons, the survey was undertaken solely in Sydney, where the four selected fieldwork suburbs—named in Table 2 above—were located in middle and outer ring locations (see Figure A7 in Appendix 1).

1.4

Report structure

The remainder of this report is structured according to the study’s three ‘overarching issues’ defined in the original research proposal (as set out in Section 1.1). Its remit is to discuss these issues in terms of overall learning from the research program and to summarise the contribution of this program. Thus, Chapter 2 reflects on the conceptualisation of socio-spatial disadvantage and on the relationship between spatially concentrated disadvantage and the functioning of cities. This draws primarily on research elements (i)–(iii) as listed in Table 2. As well as reflecting on the findings of our published 2001–11 housing market analysis (research element (iii)), this chapter is informed by an updated breakdown of disadvantaged area housing market performance 2011–14. Chapter 3 draws out study findings about the changing geography of disadvantage and its consequences, both for residents of disadvantaged places, and for metropolitan systems more broadly. This touches on the debate about the inherent implications of living in a ‘poor area’—that is ‘neighbourhood effects’. However, drawing mainly on research elements (v) and (vi), its main focus relates to the local features of the particular disadvantaged areas studied in this research and the extent to which these are seen to impact on residents’ welfare. Drawing primarily on research element (vi), but also referring to other components of the study, Chapter 4 focuses on policy responses to spatial disadvantage and to spatial concentrations of socially disadvantaged people. Finally, in Chapter 5, we reflect on the research methodology, suggest future research directions and draw some broad conclusions from the research. Much of the argument, many of the references, and some of the data in this report are new. However, bearing in mind the report’s status in rounding off its associated research program, and given that readers cannot be assumed to have read the entire series of already-published outputs (see Table 2), it necessarily contains a certain amount of material reproduced from these reports. Where this occurs it is indicated in the text.

14

CONCEPTUALISING SPATIAL DISADVANTAGE

2

‘Policy-makers cannot meet the challenge of designing policies for rectifying disadvantage without a proper philosophical and conceptual discussion of what disadvantage is.’ (Wolff & De-Shalit 2007, p.11)

2.1

Introduction

How we understand spatial disadvantage in cities is critical to understanding why and how this matters for some people, how it affects particular urban places and for the functioning of cities more generally. It is also essential for developing the types of interventions that would prevent and mitigate such disadvantage. This chapter discusses the conceptualisation of spatial disadvantage in the Australian urban context and seeks to unpack the relationship between spatially concentrated disadvantage and the functioning of cities. It distils some key points from a large volume of literature on disadvantage and spatial disadvantage in the context of urban restructuring in large cities and draws primarily on research elements (i)–(iii)—as listed in Table 2.12

2.2

Disadvantage

Before looking specifically at spatial disadvantage, it is important to develop an understanding of ‘disadvantage’, which is often used interchangeably with ‘social disadvantage’ and ‘socioeconomic disadvantage’. While many scholars and policy commentators use the term disadvantage liberally, few examine it in any detail, or define it, perhaps because ‘what it means to be disadvantaged and how to measure it are challenging and contentious issues' (McLachlan et al. 2013). A key point is that disadvantage cannot occur without advantage. It is embedded in an understanding of inequality, which recent, highly influential work suggests affects economic development and social wellbeing (e.g. Wilkinson & Pickett 2009; Piketty 2014), as indicated in Section 1.2.1. In this context, it is important to understand broad perspectives on inequality that underlie ideas about disadvantage. As proposed by Wolff and De-Shalit (2007) in an important book, Disadvantage, there are two distinct theoretical perspectives on inequality that must be reconciled: distributive inequality and social/relational inequality:  Distributive inequality refers to lack of equality in the distribution of resources which could

include such areas as wealth, income, or standard of living. It refers primarily to economic processes and material inequality.  Social inequality (or a relational view of inequality) refers to the ways in which ‘people’s lives

can be affected by government policies, the nature of the society they live in and how they treat each other’ (Wolff & De-Shalit 2007, p.6). It is about the ways in which people treat each other in ways that may not be shaped or defined by economic inequality (Wolff 2008). In this sense, social inequality could be seen in exploitation, discrimination and exclusion at both an individual and more systemic level. The challenge, as outlined by Wolff and De-Shalit (2007), is to develop a conceptual framework which integrates ideas about distributive inequality and social inequality since they are likely to be related in practice. Viewed in this way, it is possible to see the operationalisation of disadvantage through a range of concepts which have been widely used to consider aspects of distributive and social inequality, each drawing on distributional and social inequality theories and often adding their own particular ‘twist’. The examples of widely used concepts considered below not only 12

The authors wish to acknowledge the contribution of research team members Professor Andrew Jones and Dr Jonathon Corcoran (University of Queensland) and Dr Gethin Davison (University of New South Wales) in contributing to discussions about conceptualising different types of disadvantage and their measurement.

15

identify manifestations of disadvantage but also seek to illuminate the processes that produce and reproduce it.  Poverty, for example, is embedded in an understanding of distributive inequality. In most

developed societies, poverty is relative, as recognised from the time of Townsend’s (1979) seminal work, but also incorporates an understanding that some people with low incomes and a relative lack of material resources may have personal resources or family support to draw on, while for others low income will inevitably bring living standards well below community standards, encapsulated in the Australian context by the concept of deprivation (Saunders 2011).  The social in/exclusion concept is grounded in a social inequality perspective, since whether

one is included or excluded is inherently relational (a point made forcefully by Levitas 2005 in a seminal work). However, it also has the potential to consider the effects of institutional structures in excluding people as well as community attitudes (Saunders et al. 2007, p.12).  The concept of social capital as articulated by Robert Putnam (1995, 2000) and which was

highly influential 13 derives primarily from a social inequality perspective in which lack of participation in economic and social life indicates a lack of social capital. Other interpretations of the concept, however, incorporate not only social networks but access to economic as well as social and cultural resources through these networks (e.g. Bourdieu 1986).  The concept of capabilities emphasises 'the actual opportunities a person has' (Sen 2009,

p.253) and developed as a means of making quality of life assessments beyond the usual array of economic and social indicators of wellbeing. In this sense, disadvantage can be viewed as limited capabilities although, as argued by Nussbaum (2011), people may or may not take up these opportunities, they should have the freedom to choose. In the Australian context, these concepts have at various times, and to different degrees, underpinned social policy. For example, the Federal Coalition Government (1996–2007) drew on the concept of social capital in a suite of policies on ‘family and community strengthening’ (Hulse & Stone 2007). The Federal Labor Government (2007–13) used the social inclusion/exclusion concept as an overarching framework for its social policies (Hulse et al. 2011). Recognising the different perspectives on disadvantage, and the many and different concepts that have been developed and deployed, this broad field of literature suggests that disadvantage should properly be characterised as socio-economic (rather than social disadvantage), recognising both distributive and social inequality dimensions. To sum up, the key points distilled from an enormous body of literature suggest that socioeconomic disadvantage is:  A manifestation of inequality (distributional or social/relational, commonly both).  Usually relative in advanced economies (except at the margins) and thus depends on

generally held community expectations and standards, which change over time.  Includes low income but the extent to which this affects peoples’ lives will vary.  A relative lack of access to resources and opportunities (e.g. adequate education, decent

housing, employment, etc.).  Relational in that type and quality of relationships can mitigate or exacerbate disadvantage.  Multi-dimensional including economic, social, political and cultural dimensions.

13

Putnam’s (2000) book, Bowling alone: The collapse and revival of American community, was highly influential in popularising the concept of social capital and has an extraordinary 31 662 citations. Ruth Levitas’ (2005) defining book, The inclusive society? Social exclusion and New Labour, has 1750 citations which, although very high by academic standards, constitute only a fraction of the Putnam citation (Google Scholar viewed 16 February 2015).

16

 A process not just an end state: it may be temporary although can be cumulative and

compound over time. The important point for this report is that the debates about socio-economic disadvantage have been very much part of a social policy paradigm that is largely a-spatial and, somewhat distinct from a paradigm that focuses on place-based poverty (Griggs et al. 2008, p.1). In the next section, we look more specifically at disadvantage in the context of urban space and place.

2.3

Spatial disadvantage

Spatial disadvantage, like disadvantage more generally, is often taken for granted, difficult to define, and encapsulated in a number of different but related concepts which derive from different understandings of the nature of residential differentiation, as well as associated problems of measurement. In the Australian context, at least three distinct concepts of spatial disadvantage have been deployed:  Locational disadvantage, referring to the relatively poor access of residents in some places to

key resources such as employment, education, health care and public transport. Disadvantage is defined by opportunity, or lack of opportunity to access resources (Maher et al. 1992; Wulff et al. 1993; Maher 1994; Badcock 1994; Hunter & Gregory 1996; Ryan & Whelan 2010, Saunders & Wong 2014). In practical terms, this includes living in places in which there are few local jobs, a lack of schools and tertiary education institutions, poor access to GPs, specialists and hospitals, and remoteness from major cultural institutions. A key element in locational disadvantage is limited or inadequate public transport such that it is difficult, expensive and time-consuming for residents to participate in economic and social life. It is important to note that locational disadvantage can affect all people living in an area whether they themselves are disadvantaged or not. However, people with higher levels of education and resources may be able to mitigate some of the effects of locational disadvantage to a greater degree, for example, through having jobs with flexible conditions which enable some working from home, having access to private vehicles which are in good condition, etc.  Dysfunctional disadvantage, referring to areas that appear to have a disproportionate

incidence of social problems such as crime, drug addiction, unemployment, vandalism and antisocial behaviour exposure to which creates risk and fear for residents. Many of the influential explanations derive from the US and include long-standing ideas about a ‘culture of poverty’ and a ‘moral underclass' (Lewis 1966; Auletta 1982; Murray 1984; Mead 1997). Following the work of Wilson (1987), this type of thinking led to the proposal that the social and environmental effects of living in such areas had an independent effect on people’s life chances over and above the characteristics of people themselves, that is it compounded disadvantage (discussed in Galster 2012; Pawson et al. 2012—see also Section 3.4.1).  Concentrated disadvantage, referring to a disproportionate number of people with low income

and other indicators of socio-economic disadvantage (e.g. being unemployed and having a low level of education) living in particular places. Identification of such concentrations of socioeconomic disadvantage has a long history particularly in the US and European context (as reviewed in Pawson et al. 2012). In the Australian context, such concentrations have been identified using data from the ABS Census of Population and Housing as well as other ABS data (e.g. Baum et al 2006). Vinson (2007, 2009) used a combination of Census data and administrative data from government departments to extend the focus to factors that restrict opportunities and indicate lower levels of wellbeing. Locational disadvantage can be seen as mainly a manifestation of distributive inequality in its broadest context, that is the past and current distribution of resources for schools, hospitals, public transport, etc.; the concept of dysfunctional disadvantage—or places with social problems—is related to social/relational inequality while concentrated disadvantage appears to

17

include elements of both. In practice, it is likely that these different concepts of spatial disadvantage are related to each other in different ways. For example, locational disadvantage in the form of lack of access to education or difficulty in finding jobs may be a cause of low income and poverty. Concentration of people with socio-economic disadvantage has been argued in the international literature to create conditions for social problems such as lack of economic selfsufficiency, violence, drug dependency and poor educational aspiration (Wilson 1987; Vinson 2007; Galster 2012). The deployment of these concepts suggests in different ways that spatial disadvantage, like disadvantage generally, results from both distributive and social inequality, is relative, multidimensional, includes a relative lack of access to resources and opportunities as well as people with low incomes and other attributes indicating disadvantage, and is a process rather than an end state. As Saunders (2011, p.3) notes: … locations, like individuals, differ along a spectrum of disadvantage that reflects differences in local labour markets (and hence job opportunities), the availability and adequacy of local services, social and community facilities and the strength of informal networks that provide support to individuals and families in times of need or crisis. Unless research has the capacity to capture these effects, it will be incapable of identifying all of the factors that contribute to poverty, resulting in a biased view that covers only a limited range of risks. What is missing in Australian debates about spatial disadvantage is the relationship between the restructuring of urban space in Australian cities and disadvantaged places, which we discuss next.

2.4

Urban space and place

As foreshadowed in Chapter 1 (Section 1.2.1), a growing interest in the economic and social restructuring of urban space in Australia (e.g. Badcock 1984, 1994; Forster 1995; Randolph 2004b, 2006; Baum & Gleeson 2010; Randolph & Tice 2014) has occurred at the same time as an increased focus on disadvantaged places (e.g. Peel 2003; Warr 2005; Vinson 2007). It is important to have a clear, contemporary14 understanding of the difference between urban space and place, and the relationship between them, to develop a clear conceptualisation of spatial disadvantage in an urban context, as considered in detail in the following chapters of this report.

2.4.1 Space and place Although there are many different strands in the literature (in the same ways as there are for disadvantage discussed above), the essential difference between space and place is that while space is abstract and general, place is particular and specific.  Space has been conceptualised as the abstract space of modern capitalism,

15

produced by economic transactions and public policies. Space is structured by a myriad of economic transactions interacting with public policies that may configure space in different ways such as market segments and administrative units. Conceptualising space in this way adds another (i.e. spatial) dimension to an understanding of processes of distributive inequality. It draws attention to economic and political factors.

 Place is in many respects an easier and more practical concept but has been defined in at

least three different ways (Agnew 2011). First, place can be viewed as a location or a system of physical places including the infrastructure that connects them. Second, it has been seen as a series of locales where everyday activities take place. Third, it is deployed as a sense of place or identification with a unique community which is not necessarily tied to physical 14

Agnew (2011) points out that the ancient Greek philosophers, such as Plato and Aristotle, distinguished between chôra and topos with the first referring to place and the modern term space coming from the second of these. 15 The most important contributions in this area are by Henri Lefebvre, Manuel Castells and David Harvey, a view also articulated in a review by Kesteloot et al. (2009).

18

location (e.g. an ‘internet community’). Conceptualising place accords more, although not exclusively, with a social/relational view of inequality. It draws attention to physical, social and cultural factors. The relationship between space and place is a complex one. Restructuring of urban space is likely to reflect, and potentially reinforce or mitigate, distributional inequality and may not necessarily accord with residents’ perspectives of place or the ways they live their lives. As notably articulated by Tuan (1977, p.6) ‘… what begins as undifferentiated space becomes place as we get to know it better and endow it with value’. In other words, places are shaped not only by the effects of markets and public policies but also by the views and actions of the people who live there. This dynamic emerged strongly in the research program in that the residents of disadvantaged places identified in the study often viewed them in positive, relational terms while acknowledging some aspects of locational disadvantage and the effects of socio-economic disadvantage more generally (Pawson & Herath 2015).

2.4.2 Configuring urban space and place Distributive inequality is substantially created, experienced, and mitigated or exacerbated in Australia’s cities, with just under two-thirds of the population (or 15.3 million people) living in the nation’s eight capital cities (June 2013).16 The city can be seen as what Harvey (2009 [1973]), p.68) memorably termed ‘a gigantic resource system’. In addition to the obvious distributive mechanism of labour market income, there are also ‘hidden mechanisms’ of redistribution which affect the real incomes of groups of people in urban space. These mechanisms are market and policy decisions on issues including transport networks, zoning, and location of public institutions and facilities, as well as residential differentiation (Harvey 2009 [1973]), ch.2). Urban places, however, are not just a spatial working of broader economic factors in a particular location, they are also created by people, provide locales in which people live and may also, although not always, provide a ‘sense of place’. Experience of living in particular places in the city is ‘also fashioned by their individual perceptions, mental maps and spatial practices’ (Tonkiss 2005, p.13). Thus while it is important to consider the economic processes that shape and reshape cities, it is also important to examine the ‘micro-networks of social action that people create, move in, and act upon in their daily lives (Smith 2001, p.6). Just as disadvantage is embedded in distributive and social inequality, conceptualising spatial disadvantage in cities must consider both the ways in which space is configured (primarily economic and political) and the making of places (primarily social and cultural) 17 (see e.g. Fainstein 2010; Harvey 2012). Understanding spatial disadvantage requires attention to the ongoing dynamic between the two, as illustrated in the examples below.  There is a strong ‘community studies’ tradition that provides rich and detailed accounts of the

lives of people living in disadvantaged places (some notable Australian examples are Bryson & Thompson 1972; Peel 1995, 2003; Bryson & Winter 1999; Warr 2005). Such studies focus on lived experience of disadvantaged places but they are grounded in an understanding of the broad economic changes that have reshaped urban space, such as a decline in manufacturing and de-industrialisation which has adversely affected particular places. This is particularly evident in two studies of the same Melbourne suburb more than 25 years apart, in which the effects of deindustrialisation (decline in manufacturing) had a very significant impact on experience of place and local community (Bryson & Thompson 1972; Bryson & Winter 1999).

16

ABS 2014 Regional Population Growth, Australia, 2012–13, cat. no. 3218.0, ABS, Canberra, http://www.abs.gov.au/ausstats/[email protected]/Products/3218.0~2012-13~Main+Features~Main+Features?OpenDocument It is important to note that the ABS uses an ‘economic definition of cities that extends beyond the current built up area (Greater Capital Cities). 17 This is a simple characterisation to draw attention to some important factors. Of course, in practice, it will be much more complicated than this.

19

 Studies of urban renewal have highlighted that policy-makers and others may view an area,

and its residents, as highly disadvantaged and design interventions to ‘improve’ place as part of on-going reconfiguration of urban space. This may result in higher quality housing, transformation of retail businesses, changed configuration of open space and a higher level of amenity and aesthetics. Residents may not see their areas as being disadvantaged (Morrison 2003) and renewal may result in a loss of place, both physically through relocation (Minto Resident Action Group 2005; Hulse et al. 2004) and for those who stay through changes in the local social structure experienced as ‘loss of community’ (Shaw & Hagemans 2015).18 We argue that spatial disadvantage must be grounded in an understanding of distributive and social inequality as these play out in space and place; it is not sufficient to focus only on places disconnected from their broader urban context. The drivers of distributive inequality are labour markets, housing markets, commercial and other markets that interact with public policies in different ways to configure and reconfigure urban space. These processes shape the resources available in particular places (e.g. access to jobs and transport), but places are also made by distinctive patterns of living and the meaning that people attach to living in particular places in sometimes unpredictable ways.

2.4.3 Flows It should be clear from the discussion to date, that the reconfiguration of urban space is ongoing and that this process helps shape and re-shape disadvantaged places within the city. A critical perspectives paper drafted within this research program argued that following Castells (1996), it is important to see cities as not just comprising ‘space of places’ but also ‘space of flows’ in which people are not bounded by place but move around in relation to jobs, technology, transport and information (Hulse & Pinnegar 2015, p.1). New types of ‘virtual’ economic exchanges using information technology have become commonplace in cities in ways unimagined 20 years ago. As one author puts it: … in a speeded up and wired world, urban space appears as more ‘porous’, urban agents (whether people, technical actors or symbolic forms) more mobile. (Tonkiss 2005, p.149) This development has generated some important debates about whether places are less important to people than previously. First, does the rise of virtual flows rather than physical ones reinforce the concentration of economic activity in cities rather than disperse it, thus compounding competition for urban resources and generating greater inequality, as suggested by Coyle (1997, p.192)? Second, does the availability of electronic communication in what has been called the ‘weightless world’ (Quah 1996) reinforce distributive inequality as some people lack access to such technology (the ‘digital divide’), thus exacerbating social inequality, since while some urban residents use such technology to enhance their ’social connectedness’, the lives of others remain bounded by place? 19 Third, is place less important in an era of electronic information and communication technologies enabling people to connect with others with similar interests and to adopt common lifestyles irrespective of physical location, as indicated by Castells (1996)? We can conclude from this that it is important not to assume a priori either that disadvantaged places are fixed or that people are ‘stuck’ in place, but rather to examine mobility as a driver of change in cities (Creswell (2010, p.551). This can include personal mobility but also residential mobility. As highlighted by Pearson & Lawless (2012, p.2034) in their assessment of residential mobility in regeneration areas in England’s New Deal for Communities Programme: Mobility is an essential component in appreciating the dynamics of neighbourhoods, and their rationale within wider geographic contexts. 18

In the international literature, Allen (2008) provides perhaps the most compelling account of the tension between restructuring urban space and place as a locale for rich social relationships between working class residents who ascribe different meaning to ‘their place’. 19 Castells (1996)) in his work recognised that places where people lived were still very important for some people.

20

More than a third of Australian households changed their place of residence between 2006 and 2011 (ABS 2012). These moves can have significant impacts on the ‘sending’ and ‘receiving’ localities, and on the functional relationships between them. Mobility as channelled by housing markets plays a role in spatial polarisation by socio-economic status of residents in Australian cities (Randolph & Holloway 2005b; Wulff & Reynolds 2010; Randolph et al. 2010; Hulse et al. 2010). Monitoring the flows in this way is a complex task (see Robson et al. 2008). Research into this dynamic in Sydney finds that while flows to some outer areas have increased, in other cases, mobility has become restricted to shorter ‘within area’ moves. This suggests that in some regions with a cluster of disadvantaged suburbs, connection to metropolitan housing markets may be weakening (Randolph et al. 2010; Pinnegar & Randolph 2013). No discussion of population flows in the Australian context is complete without an understanding of the critical role of international migration, and migrant settlement patterns. In 2011, almost a third of Australians had been born overseas (ABS 2012), reflecting different waves of migration over the post-war years (DIMIA 2009). This context differs from some of the international literature where spatial disadvantage is associated with regional economic and population decline. There are clear ‘entry point’ suburbs in Australia’s largest cities and these have traditionally provided both a ‘quick springboard’ and a longer term ‘step up’ function for new migrants. The issues, as highlighted by Hulse & Pinnegar (2015, pp.18–19) are whether traditional migrant settlement patterns are becoming blocked, in part because of housing market dynamics. They point in particular to the role of the private rental sector in accommodating much of the initial, and some longer term, demand from migrants and humanitarian refugees and the association between high levels of private renting and spatial disadvantage. Some associated issues were brought to light in this research via the case study work in Auburn (NSW) and Springvale (Vic), as reported by Cheshire et al. (2014). The importance of conceptualising flows is to understand that while governments might focus on infrastructure and services which help re-configure space, and companies might concentrate on the most profitable areas to locate their businesses, people live in specific places which not only provide differential access to resources such as jobs and transport but also create place through their everyday activities. In conceptualising disadvantage, it is as important to consider the ways in which mobility affects disadvantage and spatial disadvantage, particularly when it is not chosen, as much as it is to investigate the extent to which people are ‘stuck’ in particular places.

2.5

The importance of scale

A key factor in conceptualising spatial disadvantage is the appropriate scale. This could range from a few houses or streets, a ‘neighbourhood’, a suburb, an administrative unit (e.g. a public housing estate, a Census-defined unit to a postcode or a local government area), an urban region or a city. Which scale is selected can make a difference to the designation of particular locations as disadvantaged, conclusions drawn and policy interventions designed.

2.5.1 Dysfunctional disadvantage experienced at a small scale Spatial disadvantage defined in terms of places with social problems is typically experienced at a very fine scale—a few streets or a neighbourhood. In the survey of residents conducted for this project, the problems identified were mainly very local—such as anti-social behaviour by nearby residents, graffiti and littering and experience of ‘car hooning’. Consistent with other Australian research, our findings suggest that place matters in terms of whether people feel safe in their neighbourhood and their experiences of the local environment (Neighbourhood Renewal Unit 2008). Places with social problems are not necessarily related to concentration of disadvantage. A few people or families can create problems which make for an unpleasant and unsafe local environment, potentially affecting many residents.

21

2.5.2 Locational disadvantage in a city-wide or city region context Locational disadvantage is usually defined in terms of a large geographic scale. At its broadest level, the city as an economic unit that ‘represents the city in a wider sense’ (ABS 2006, p.15) and extends beyond the current contiguous built up area of the city to include those areas that are economically connected to the city. 20 This concept proved extremely valuable in the research program in identifying concentrations of disadvantaged people in areas beyond the built up area of the city, indicating the extent of displacement of lower income people outward from the urban core. Urban regions are also of value in considering the ways in which distributive inequality plays out through the restructuring of urban space. For example, if transport and infrastructure resources are worse in one region rather than in another, this will affect its relative attractiveness for private investment in retailing, commercial, industrial and investment in residential property. Such investment will affect the degree of locational disadvantage (and advantage) and, through housing markets, the socio-economic characteristics of residents.

2.5.3 Concentration of socio-economic disadvantage: a problem of scale The literature on concentration of disadvantage typically focused on the ‘neighbourhood’ scale, particularly in research emanating from the US and the UK. What constitutes ‘neighbourhood’ in low density Australian suburbs, characterised by detached housing and some small apartment buildings, is likely to differ from the high density neighbourhoods of inner city Detroit, Chicago or London. Arguably, neighbourhoods in Australian cities are more porous and less well defined. High car ownership in middle and outer suburbs makes available a greater range of locales for everyday activities. Of course, for those lacking a car or unable to drive due to health or older age, these types of locations may result in social isolation, although the degree to which this is experienced will depend on the strength of local social relations (Baum & Gleeson 2010). In this research program we argued that ‘suburb’ is a more appropriate scale for considering concentration of people who are socio-economically disadvantaged. Suburbs have identifiable physical boundaries and have a commonly understood meaning, as evidenced by real estate advertisements, media coverage, and other narratives of place. While the suburb is not an ABS geography, it is possible to aggregate ABS spatial units to approximate suburbs.21 It is possible to use ABS administrative units of varying sizes (ranging from the very small to a broad view of the city). These have the advantage of enabling analysis of data at various scales. For example, Baum et al. (2006) in their study of advantage and disadvantage used Statistical Local Areas or SLAs which are larger than suburbs and approximate local government areas or parts of local government areas but do not necessarily have any inherent meaning for residents. Some of the other administrative units may have advantages in terms of availability of administrative data, such as postcodes (used by Vinson 2007) but also may have no inherent meaning for residents. In brief, while it appears that locational disadvantage is the result of city-wide factors experienced locally, and the effects of places with social problems are best considered in the ‘local’ setting, the appropriate scale for identifying spatial disadvantage as a concentration of socio-economic disadvantage is more difficult to assess. Most of the US and Canadian literature in this field uses the ‘census tract’ (CT) as its prime scale of analysis (Seguin et al. 2012). Although variable in

20

The spatial unit selected for the research program was the Capital City Statistical Division (SD) as defined by the Australian Bureau of Statistics as likely to accommodate anticipated population growth and city development within the boundary over at least 15–20 years (ABS 2006, p.15). 21 The State Suburb (SSC) is a Census-specific area where Statistical Areas Level 1 (SA1s) are aggregated to approximate suburbs (ABS 2011 Census Dictionary, 2011, cat. no. 2901.0, ABS, Canberra, viewed 24 February 2015, http://www.abs.gov.au/ausstats/[email protected]/Lookup/2901.0Chapter23102011#SSC.

22

size, these contain, on average, around 1600 dwellings and 4000 people (both in Canada and the US).22 In the UK, neighbourhood studies have often focused on electoral wards (averaging 5500 population), although the profusion of outputs from the recent New Deal for Communities evaluation (e.g. Lawless 2006; Pearson & Lawless 2012) related to program areas typically encompassing 9900 people. Australian suburbs (SSCs) are in this range, although they vary in size between cities (containing between 1500 and 2600 dwellings and between 4000 and 7300 people.23 In informing area-based intervention policy, Canadian researchers have recently argued that both macro-scale (census district) and micro-scale (census ‘dissemination area’) analysis may be more appropriate than the traditional meso-scale CT level emphasis (Seguin et al. 2012). Part of the problem highlighted here is that poverty ‘micro-zones’ may be hidden within CTs. Equally, in designing policy measures to address disadvantage—such as to reduce excessive early school leaver rates—there is a need to recognise that ‘the district level [average population 53 000] better corresponds to the area in which [target group members] live on a daily basis than the dissemination area [average population 600] which is very small’ (Seguin et al. 2012, p.244). As an illustration of the importance of analytical scale in the Australian context, public housing estates are often seen as ‘concentrations of disadvantage’. If this scale is selected (or more usually an aggregation of small ABS spatial units approximating public housing estates), it would appear that there is indeed a concentration of social-economically disadvantaged people. However, as argued by Darcy (2010, p.13), this is largely tautological since public housing residents are selected on the basis of the most extreme indicators of socio-economic disadvantage. Identification of ‘concentration of disadvantage’ at a very small scale does not indicate economic and social context. In Melbourne, for example, some public housing is located in what have become relatively advantaged places due to the gentrification of inner suburbs, with good transport and access to jobs, services and facilities. Nevertheless, there may well be other types of difficulties for residents of such areas including a local predominance of ‘high end’ retailing, services and facilities targeted at more affluent residents and community organisations reflecting a gentrified demographic. These factors may create difficulties in day-to-day living in advantaged places. On the other hand, much Australian public housing is situated within wider localities likewise containing concentrated socio-economic disadvantage, and which may also experience locational disadvantage. Thus whether there are ‘clusters’ of disadvantage at a larger geographic scale, and the extent of locational disadvantage, help shape opportunities and lives of residents.

2.5.4 Some issues It is important in considering scale for identification of places with a concentration of people who are socio-economically disadvantaged to be aware of the dangers of 'ecological fallacy’, that is not assuming that everyone in a disadvantaged place has characteristics of socio-economic disadvantage, which clearly they do not. This has been described as typifying: … urban areas in terms of their worst features, overlooking the finer grain of interaction and organization which underlies them. …. This fallacy is one that labels individuals or communities as much as places. (Tonkiss 2005, p.50) The extent to which place matters in terms of the locales in which people live their lives is an open question and cannot be predetermined by the scale of analysis of spatial disadvantage. Australian 22

The US Bureau of Census defines a Census tract as: ‘A small, relatively permanent statistical subdivision of a county delineated by a local committee of census data users for the purpose of presenting data. Census tracts nest within counties, and their boundaries normally follow visible features, but may follow legal geography boundaries and other non-visible features in some instances, Census tracts ideally contain about 4000 people and 1600 housing units.’ US Bureau of Census, Census Glossary, viewed 16 February 2015, http://www.census.gov/glossary/#term_Censustract. 23 The average number of dwellings and people per suburb is higher in Melbourne than in Sydney or Brisbane.

23

studies have found that although residents may be further disadvantaged by the social problems manifest in some places they are not necessarily bound by local neighbourhoods in terms of their daily lives (Peel 1995, 2003; Warr 2005) although all may be affected by the pernicious effects of stigma (Jacobs & Flanagan 2013). UK research on low-income neighbourhoods has also found that people’s daily routines, although very individual, extended beyond the neighbourhood for work, education, shopping, leisure, access to public services and social networks (Robinson 2011, p.129). Finally, although areas with a concentration of disadvantaged people are often assumed to create adverse effects for residents, empirical research suggests that residents in disadvantaged places access support from family and friendship networks that are not restricted to place (Robinson 2011) and may in fact have certain advantages in terms of mutual support and practical assistance in ‘hard times’. According to one study, neighbours can matter as means of ‘getting by' and dealing with the challenges of poverty and disadvantage (Bashir et al. 2011, pp.20–22). Issues around analytical scale, as these relate to our own research findings, are further discussed in Section 3.4.2.

2.6

Drivers of spatial disadvantage

The restructuring of urban space in Australian cities has been associated with the suburbanisation of poverty. The inner city areas which were seen as areas of spatial disadvantage up to the late 1970s (Kendig 1979; Logan 1985) have been extensively gentrified. Poorer residents were displaced and dispersed to some middle and many outer suburbs (Randolph & Holloway 2005b; Forster 2004; Baum et al. 2006; Gleeson & Randolph 2002; Randolph & Tice 2014). The disadvantaged places identified in this research (see Chapter 3) were all in middle and outer suburban and fringe urban locations. In this section, we examine the key drivers of spatial disadvantage in Australian cities: urban governance and labour markets and housing markets.

2.6.1 Urban governance Spatial disadvantage does not occur in a vacuum. The ways in which cities are planned and managed, and the institutional settings that are established, provide the context for restructuring of cities and the shaping of areas of relative disadvantage (Burke & Hulse 2015, p.12). In what Burke and Hulse (2015, p.16) in a Critical Perspectives Paper written for this project refer to as ‘the hidden hand’ of public policy, policies in a broad range of areas (including income support, taxation, labour markets, industry, planning, transport, immigration, education, health and housing) have spatial outcomes, even where these are not explicitly considered. The location and type of physical and social/cultural infrastructure creates places that are relatively resource rich or poor. For example, the concentration of cultural institutions, sporting venues and universities in inner cities creates resource-rich areas which attract residents who value these. Conversely, a lack of transportation infrastructure between and within outer suburbs may make these areas relatively unattractive. Such policy decisions have flow-on effects on private investment decisions (e.g. on the location of industry, commerce and retail and on residential property prices). When spatial disadvantage in Australian cities was concentrated in inner areas (Kendig 1979), funding in previous eras had provided good access to schools, libraries, public transport and a whole range of other facilities and services. With the suburbanisation of disadvantage, discussed above, funding structures and flows associated with various policy domains became more problematical. A key issue, therefore, is the extent to which funding flows have followed the suburbanisation of disadvantage logic.24

24

In considering this question, it is important to note that key services such as health, education and housing are funded by federal and state/territory governments in Australia, unlike the situation in the US where such services are heavily dependent on local funding.

24

Tracing funding flows at a spatial level is a complex exercise. A detailed examination of federal and state housing subsidy flows in Melbourne for this research program (Groenhart 2014) found compelling evidence that housing subsidies flow disproportionately to advantaged not disadvantaged postcode areas. Expenditure on direct housing assistance (public housing and Rent Assistance for low-income private renters) was well targeted to disadvantaged people living in disadvantaged areas. 25 The benefits of indirect housing assistance through the tax system flowing to advantaged areas far outweigh the direct financial benefits targeted at socioeconomically disadvantaged areas (Groenhart 2014, pp.1–2). Metropolitan planning, on the other hand, is explicitly spatial. It has provided the institutional framework for the development of Australia’s largest cities. This is a responsibility of state governments, although aspects are implemented by local government. Metropolitan planning has largely been based on a model of the mono-centric city (with Sydney a part exception) with public and private transport infrastructure designed to take people to the city centre. Not surprisingly, this has reinforced the concentration of higher quality jobs in inner suburbs and revitalised inner urban housing markets (see Section 3.3.5). For many people living in disadvantaged urban places in middle or increasingly outer suburbs, while they can travel to city centres, they cannot travel more locally to centres of employment, at least not without one and sometimes two cars, as we discuss next.

2.6.2 Labour markets Australia’s urban economy has restructured away from manufacturing and towards services, particularly financial services and education (Beer & Forster 2002). This has played out spatially in the increased concentration of higher occupational status jobs in services industries in inner city areas and loss of lower occupational status jobs in manufacturing, utilities and construction in these areas. On the other hand, in outer urban areas, high skill manufacturing jobs have been replaced with routine production jobs (Dodson 2005). Such jobs are typically lower occupation status and have been particularly vulnerable to broader global and national economic changes (Baum et al. 1999). In some cases, the local economic base has been decimated through these changes leading to what Baum & Mitchell (2009) identify as ‘red alert suburbs’. While this is a story common to many advanced countries, Burke & Hulse (2015, pp.9–10) argue that the depth of spatial disadvantage resulting from these changes is not as great as in deindustrialised areas of cities such as Baltimore, Belfast, Buffalo, Detroit, Manchester, Pittsburgh and Sheffield in which a major loss of jobs was associated with population decline, falling house prices and property abandonment (Power et al. 2010; Brookings Institution 2007). The labour markets of Australian inner suburbs transformed, as indicated above, with an intensification of higher occupational status jobs. The workers’ cottages of inner suburbs were likewise transformed through renovation, extension and demolition/rebuild into highly desirable and higher price housing. Like many commentators in the early 1980s, Badcock (1984) saw the labour market as the key driver of disadvantage, somewhat mediated by spatial location. He asserted: The relative location of those disadvantaged families living on fringe housing estates in Australian cities is not the root cause of their inequality. That derives from their position in the labour market, which in turn limits where they will be located by the housing market. (Badcock 1984, p.53) While this may well have been true 30 years ago, and labour markets remain important determinants of people’s access to jobs, occupational status and incomes, housing markets do 25

Households in the most advantaged 25 per cent of postcodes in Melbourne received on average $4600 in direct and indirect housing benefits while those in the most disadvantaged 25 per cent received on average $2800 in such benefits in 2011–12 (Groenhart 2014, p.2). This work provides a spatial dimension to earlier work by Yates (2009, p.6) who found that 'taxes in relation to housing provided most benefit to those who least need it’, namely older owner occupiers in the top income groups.

25

not just mediate the effects of labour markets but have become important, independent drivers of distributive inequality and spatial disadvantage, as we discuss next.

2.6.3 Housing market restructuring Housing markets shape where people live. Housing market factors, including access to economic resources (income and wealth) and the price/rent of housing are critical in shaping access to place. As discussed earlier in this section, housing markets are in turn affected by institutional settings and a range of government policies. They are also shaped by a series of social and cultural factors that shape individual trade-offs, for example, between housing type, housing tenure and location (Hulse et al. 2011, p.18). Deregulation of the financial system in the mid-1980s in conjunction with a shift to independent monetary policy and the consequent fall in nominal interest rates (Yates & Yanotti 2015) provided the environment for increased availability of housing finance over the subsequent three decades. Increased household incomes due to continuous economic growth since the 1990s alongside a rise in two-income households and a general investment preference for ‘bricks and mortar’ has provided a high level of demand for residential property from both owner occupiers and investor landlords. The combination of available and affordable mortgage funds and a high level of demand, in the context of a constrained supply of housing, put upward pressure on residential housing markets. The flow of funds went disproportionately into residential property in inner suburbs and some middle suburbs which were relatively advantaged in terms of public and private infrastructure and resources. The first effect was to increase house prices, and therefore rents, in these areas to a greater degree than many outer suburbs (Burke & Hulse 2015, p.18). The second was that the relative flat bid rent curve (Alonso 1964) of the early 1980s had become a classic steep curve by 2011 with higher prices the closer to the inner city (Hulse et al. 2010, pp.73–79). For those owning property in inner and many middle suburbs, this process dramatically increased their household wealth. The implications for people on lower incomes, however, have been stark. Whereas in the early 1980s they could afford to buy or rent housing in a variety of locations, by 2011 they had been priced out of much of the metropolitan market, despite increases in real incomes even at the lower deciles. The only options to buy were in lower priced (generally) outer suburbs. The research program showed that disadvantaged suburbs attracted purchasers, sometimes from some considerable distance because they wanted to get a ‘foot on the home ownership ladder’. However, many of these indicated that they wanted to move out as soon as they were able. No account of housing market restructuring would be complete without considering the rise of investment in rental housing in Australian cities, with a consequent increase in private rental households. Between 1986 and 2011, the number of rented dwellings grew by more than twice the rate of occupied private dwellings generally (Hulse et al. 2014b, p.18). While there has been some increase in investment in higher rent dwellings, the research program showed evidence of increased levels of rental investment in disadvantaged suburbs. This can be explained by expectation of an increase in capital value (a realistic expectation in the period 2001–11) and better gross rental yields in such suburbs than in higher priced areas. Disadvantaged suburbs in Australia are characterised by much higher levels of rental than their cities generally (Hulse et al. 2014a). The research showed that while owner occupiers get a considerable discount by buying in disadvantaged areas, private renters do not get a similar discount on city median rentals (Hulse et al. 2014a, p.64). In addition, relatively weak regulation of residential tenancies by international standards (Hulse & Milligan 2014) means that they are exposed to unpredictable and potentially unaffordable rent increases. To round out the picture, public housing, and subsequently other types of social housing, in Australia were well spread through the suburbs in the post-war period rather than concentrated in inner cities as in some US cities (Burke & Hulse 2015, pp.14–16). As indicated above, some of 26

these developments are now ‘islands of disadvantage’ in highly advantaged and high price housing markets and others remain well integrated into their suburbs. The research showed that some in outer suburban locations were isolated from local housing markets, with little inward or outward mobility. They appear to be places where all three concepts of spatial disadvantage might occur and overlap, compounding the socio-economic disadvantage already experienced by residents. The research program identified some such areas, mainly in Sydney, which are relatively isolated from transport and facilities. They are characterised by poor labour market opportunities, housing markets disconnected from the rest of the city and by low rates of residential mobility.26 To sum up, the rise in housing prices, and the increasing gap between well-resourced inner/middle suburbs and outer ones has not only exacerbated the longstanding asset divide based on tenure but has been an ‘engine of inequality’ between home owners in high and lower priced locations. As expressed by Hulse & Pinnegar (2015, p.20): For lower income purchasers there are concerns tied to affordability, the rise of negative equity, dependency on the car and long commutes to more vulnerable jobs in the postindustrial labour market. For low-income private renters, they are likely to reflect on-going affordability constraints and the realities for many in this tenure, and the insecurity that accompanies it, is likely to be where they remain for the longer term. Such factors may, and often do, coalesce around households disadvantaged across a range of dimensions, but the ‘concentrations’ seen in Australia are characterised by their breadth rather than depth. In contrast to some US and UK accounts which focus on the lack of residential mobility associated with disadvantaged suburbs—‘being trapped’ in low-value housing markets—the Australian evidence suggests that high levels of residential mobility can also be a manifestation of disadvantage (Fincher & Iveson 2008, p.34). A recent study of the housing pathways of lowincome Australians identified four typical mobilities: frequent moving in the private rental sector; pathways into homelessness; loss of home ownership; and the ‘revolving door syndrome’ among some social housing tenants (Wiesel 2014). The author concludes that: Spatial disadvantage can mean being trapped in a certain type of place. It can also mean being trapped in a certain pattern of mobility: often a relentless need to move, often forced and often involving deterioration (or at the very least stagnation) in living conditions. (Wiesel 2014, p.331)

2.7

Summary

It is important to have a clear understanding of spatial disadvantage, to inform policy debates in Australia. We have argued in this chapter that disadvantage27 has been operationalised through a series of concepts (e.g. poverty, deprivation, social exclusion, social capital and capabilities) that attempt to illuminate different aspects of distributive inequality and/or social inequality. Spatial disadvantage28 is operationalised through at least three concepts: locational disadvantage; dysfunctional disadvantage or places with social problems; and concentrated disadvantage, the latter referring to places that house a disproportionate number of socio-economically disadvantaged people. Spatial disadvantage, like disadvantage generally, results from both distributive and social inequality, is relative, multi-dimensional, involves people not only with low 26

In the research, these were seen as 'isolate areas’ in housing market terms, as discussed further in Chapter 3, Section 3.2. 27 We consider this best described as either disadvantage, or socio-economic disadvantage, to incorporate dimensions of distributional and social inequality; it applies to people rather than places. 28 We consider this best described as spatial disadvantage, as including both economic and social aspects rather than ‘socio-spatial disadvantage’ which downplays economic drivers and impacts.

27

incomes but also includes a relative lack of access to resources and opportunities and is a process rather than an end state. In understanding spatial disadvantage, it is important to unpack the relationship between the restructuring of urban space and the making of particular places. The restructuring of urban space through market transactions and public policies shapes relative spatial disadvantage but places are also made by people as they connect with other people and organisations, select the locales for their everyday activities, and ascribe meanings to place. It is also important to consider cities as ‘flows’ as well as ‘stocks’; the ways in which people move about the city and the extent of residential mobility are important in the dynamic process of making and remaking places with relative degrees of advantage and disadvantage. Selection of scale is critical in identifying spatial disadvantage in Australian cities. Places with social problems are best seen at a small, fine spatial scale—perhaps the neighborhood. Locational disadvantage is best viewed at a larger scale although it may have very local effects. Spatially concentrated disadvantage is more problematic since choice of scale influences whether a concentration of socio-economically disadvantaged people is in fact identified. We argue firstly that the neighborhood is too small a scale, with the suburb being preferred and that contiguity of areas of concentrated disadvantage is a key factor rather than identification of isolated small geographic areas, in view of evidence that, in Australia at least, residents’ lives are not contained by the places they live in. Spatial disadvantage in Australia has changed as a result of restructuring of urban space over the last three decades. The drivers have been urban governance and labour market restructuring that has intensified investment in inner city locations, which are now highly gentrified and resource rich. During this period, housing markets restructuring has not only reflected changes in labour markets and mediated this through place, but such markets now operate an independent effect on distributional inequality through increased disparities in wealth. The suburbanisation of poverty is characterised by home purchasers attempting to build housing wealth and finding somewhere to live (although not all do this) and renters looking for somewhere cheaper to rent, often moving around due to the nature of the dominant private rental sector, and not able to raise a deposit to build housing wealth. Finally, spatial disadvantage takes different forms and not all disadvantaged places are the same. They play a variety of roles within their broader spatial contexts. This diversity needs to be acknowledged, and these different drivers and contexts necessarily flow through to different policy responses. In the next chapter, we outline the development of a typology of disadvantaged suburbs and its implications for an understanding of spatial disadvantage.

28

3

THE CHANGING GEOGRAPHY OF DISADVANTAGE: INDIVIDUAL AND URBAN SYSTEM IMPACTS

3.1

Chapter overview and sources

Taking its cue from the second of our overarching research themes (see Section 1.1), this chapter explores ‘individual and societal impacts’ in relation to two aspects of ‘place-based disadvantage’ in urban Australia. First, it is concerned with the extent to which the spatial clustering of poor or otherwise disadvantaged people can generate a ‘neighbourhood effect’ that is ‘local to the place’ and that compounds the impacts of such disadvantage for local residents. Second, it focuses on the spatial distribution of Australia’s disadvantaged urban populations and on the ways that broader location of ‘concentration areas’ may influence both the welfare of residents and the economic productivity of cities. In considering these issues, the chapter seeks to integrate evidence from our research with broader debates in urban studies, involving both Australian and overseas contributions and perspectives. It must be acknowledged that we are grappling here with complex and multi-faceted questions and that in addressing such issues much of the ‘evidence’ to hand—including our own—must be regarded as more ‘indicative’ than definitive in nature. Over and above references to our existing initial literature review (Pawson et al. 2012), the chapter also features a range of contributions published while the current study was under way and identified through a fresh (2014/15) appraisal of ongoing debates. In framing the above discussion, however, there is a prior need to review and summarise our research findings as regards the spatial patterns of socio-economic disadvantage in Sydney, Melbourne and Brisbane, to analyse the ongoing dynamic of change, and to discuss factors underlying this dynamic. Thus, Section 3.2 and Section 3.3 provide a contextual backdrop to the heart of the chapter embodied by Section 3.4 and Section 3.5. The chapter draws on a number of reports emanating from the project and already published in the AHURI series shown in Table 2 (Chapter 1). In particular, the first part of the chapter summarises findings from our statistical analysis and mapping of socio-economic disadvantage (Hulse et al. 2014a), extended to include additional ‘change over time’ analysis. Also included in this part of the chapter are relevant key findings from our survey of residents in disadvantaged suburbs (Pawson & Herath2015). Our main source of primary evidence for the second part of the chapter is the local area studies of six disadvantaged places, as reported in Cheshire et al. (2014). The chapter is therefore structured as follows. First, we briefly summarise the findings of our previously published analysis (as in Hulse et al. 2014a) on the contemporary geography of socioeconomic disadvantage in Sydney, Melbourne and Brisbane. Next, reporting further work not previously published in the existing AHURI report series, we analyse changes over time in the spatial distribution of socio-economic disadvantage in the three cities. Reflecting upon these empirical findings, the final two sections of the chapter then discuss the two issues cited above, namely the notion of ‘neighbourhood effects’ and the wider question of how the spatial distribution of disadvantaged populations impacts on the welfare of local residents and on the social and economic functioning of cities.

3.2

Analysing the geography of socio-economic disadvantage in Sydney, Melbourne and Brisbane

As explained in Section 1.4.2, a foundation for our primary fieldwork was a detailed statistical analysis of census data to first identify, and then classify spatial concentrations of socio-economic disadvantage. Applying the methodology summarised in the above-mentioned section, this

29

generated a cohort of 177 suburbs, some 10 per cent of the three-city total, and accommodating 16 per cent of combined city populations—see Table 4.29 In all three cities, disadvantaged CDs and suburbs were highly clustered.30 Among the latter, only 25 (or 14%) were ‘isolated’ (i.e. non-contiguous with another such suburb). Again, in common across the three cities, the identified areas were primarily located in middle and outer metropolitan areas—see Figures A1–A3 in Appendix 1. Table 4: Calibrating the 2006 geography of disadvantage in Sydney, Melbourne and Brisbane Sydney

Melbourne

Brisbane

All

Total number of Collection Districts (CDs)

6,697

6,176

2,992

15,865

Number of disadvantaged CDs

1,123

892

422

2,437

17

14

14

15

815

492

418

1,725

91

50

36

177

11

10

9

10

807

684

259

1,750

72

77

61

72

0.74

0.59

0.21

1.54

18

17

12

16

% of disadvantaged CDs Number of suburbs Number of disadvantaged suburbs

3

% of disadvantaged suburbs Number of disadvantaged CDs in disadvantaged suburbs % of disadvantaged CDs in disadvantaged suburbs Population in disadvantaged suburbs (million) Population in disadvantaged suburbs as % of total city population

Source: Reproduced from Hulse et al. 2014a Notes: 1. Reflecting the timing of our analysis, disadvantaged suburbs were identified in relation to the ABS 2006 census data. 2. Collection Districts were the smallest spatial unit used in the 2006 ABS Census (average population (Sydney)—612).

Importantly, the geography of disadvantage, as calibrated using our methodology, was only to a limited extent affected by the distribution of social housing. Across the 177 suburbs classed ‘disadvantaged’ only 10 per cent of households were accommodated in social housing in 2011.31 Next, using methodology briefly summarised in Section 1.4.2, the research went on to develop a typology of the 177 ‘disadvantaged suburbs’. Differentiated according to socio-economic indicators, four distinct types of disadvantaged suburbs emerged from this analysis, although only in Sydney were all four categories represented. The four categories were also spatially distinctive in their distributions—see Figures A4–A6 in Appendix 1. Building further on the above foundation, as briefly explained in Section 1.4.3, we went on to undertake an in-depth analysis of 2001–11 housing market performance in Sydney, Melbourne and Brisbane, differentiating between ‘disadvantaged’ and ‘not disadvantaged’ suburbs, and between the four distinct ‘disadvantaged suburb’ types. This analysis revealed a strong symmetry 29

Under a variant analysis, using a ‘lowest SEIFA decile’ rather than ‘lowest SEIFA quintile’ threshold, 68 of the 177 suburbs were classed ‘disadvantaged’. This demonstrates the number of problematic localities remained substantial, even when subject to a ‘more rigorous’ definition. Therefore, while it is certainly true that there are many highly disadvantaged localities in regional and remote Australia (especially districts with substantial Indigenous populations), highly disadvantaged areas are numerous in major metropolitan centres. 30 The ‘Moran’s I’ index scores for disadvantaged CDs in disadvantaged suburbs ranged from 0.46 in Brisbane to 0.59 in Melbourne (where +1 indicates perfect positive spatial autocorrelation, 0 indicates a random pattern, and -1 indicates perfect negative spatial autocorrelation). 31 Although this is twice the average incidence of social housing in Australia (5% of occupied private dwellings).

30

between the four socio-economically differentiated ‘archetypes’ and distinctive local housing market structures and performance metrics. Table 5 below summarises the key distinguishing socio-economic features of each defined typology category. ‘Distinguishing features’ here are the respects in which each archetype stood out from the norm for all disadvantaged suburbs. Also shown in the table are the housing marketrelated labels subsequently ‘mapped onto’ the socio-economically determined categories. Table 5: Disadvantaged suburbs typology in socio-economic and housing market terms Distinguishing socio-economic characteristics

Housing market designation

Type 1

High on young people and single parent households

‘Isolate suburbs’—High social rental; median sales prices and rents far below city-wide norms

Type 2

High on overseas movers, high on twoparent families

‘Lower price suburbs’—Relatively affordable house prices and distinct low rent market

Type 3

High on residential mobility but low on overseas movers, high on older people

‘Marginal suburbs’—Markets detached by distance from mainstream markets; high concentration of low sales prices and rents

Type 4

High on overseas movers, high on reduced unemployment and incidence of low status jobs

‘Dynamic improver suburbs’—Sales prices and rents moving rapidly towards city-wide norms

Source: Reproduced from Hulse et al. 2014a

3.3

The changing spatial pattern of socio-economic disadvantage

3.3.1 Overview As noted in Section 1.4.2, in identifying ‘disadvantaged suburbs’ our central analysis was necessarily reliant on the 2006 SEIFA index. Subsequently, however, it has become possible to make reference to 2011-based SEIFA rankings. This facilitates a comparison that suggests revealing trends over time.32 Unlike our research findings summarised in Section 3.2, this analysis has not been previously published in the AHURI report series listed in Table 2. Here, we are building on an existing literature calibrating the ‘suburbanisation of disadvantage’ in Australia’s major metropolitan areas. As appraised in detail in our earlier literature review (Pawson et al. 2012), this has tracked a gradual breakdown in the predominantly inner city distribution of urban poverty and deprivation as it existed in the 1970s (see also Burke & Hulse 2015). By the late 1990s it was noted that ‘unlike US and British cities where unemployment, poverty and urban decay is overwhelmingly concentrated in the inner city, the spatial distribution of the poor in our big cities is [becoming] much more ambiguous’ (Badcock 1997, p.246). And by 2001, the Sydney census tracts with the most severe disadvantage were ‘overwhelmingly concentrated in the middle and, to a lesser extent, outer suburbs’ (Randolph & Holloway 2005b, p.57).

3.3.2 The growing spatial concentration of socio-economic disadvantage As shown in Table 6 below, the overall number of disadvantaged suburbs increased from 177 to 187 between 2006 and 2011, although this was largely accounted for by the 10 per cent increase recorded for Sydney. Similarly, in Sydney, suburbs classed as ‘disadvantaged’ increased their 32

It should be noted that, while ABS cautions against the comparison of SEIFA scores between censuses, our approach refers to rankings rather than scores. It is also acknowledged that the new ABS census geography adopted in 2011 means that comparison with 2006-based analyses cannot necessarily be undertaken on a wholly ‘like for like’ basis. In focusing on the smallest areal unit at which census data is published, our 2011 analysis related to ‘SA1’ areas (average 2011 population in Sydney: 421) rather than Collection Districts—CDs (average 2006 population in Sydney: 612).

31

share of all suburbs by two percentage points to 13 per cent. In all three cities the period apparently saw growing spatial concentration of disadvantage as reflected by a rising share of disadvantaged CDs (SA1s in 2011—see footnote 32 below) contained in disadvantaged suburbs. Collectively, as shown in Table 6, this proportion increased from 72 per cent to 74 per cent. Consistent with the above finding is the disproportionate increase in the number of ‘disadvantaged suburbs’ identified under the more rigorous SEIFA decile threshold (i.e. where the label ‘disadvantaged’ is applied to the lowest 10% of the national SEIFA distribution, rather than the lowest 20%). As shown in Table 6, there were 68 such suburbs in 2006—38 per cent of the disadvantaged suburbs identified under the quintile threshold. Not only did this number rise to 83 suburbs in 2011, but as a proportion of those enumerated under the quintile threshold these accounted for 44 per cent. Therefore, while it remains the case that the most highly disadvantaged suburbs are slightly underrepresented within our disadvantaged suburb cohort, (accounting for less than 50% of these), there appears to be a trend towards an even split between Sydney/Melbourne/Brisbane suburbs identified with respect to the first and second deciles of the SEIFA distribution. These findings appear to indicate an ongoing process of spatial concentration in the geographic distribution of socio-economic disadvantage.33 This tendency is consistent with earlier research findings by Gregory & Hunter (1995), recently updated by Kelly & Mares (2013). These analyses have shown how, since the 1970s, there has been a spatial polarisation of economic participation such that unemployment has become more concentrated in lower status suburbs (ibid, Figures 25 and 26). Possibly linked with this trend has been the tendency for Australia’s suburbs to become less socially mixed over time (Taylor & Watling 2011). As seen by Kelly and Mares, these developments indicate a tendency for the increasing spatial polarisation of Australia’s cities. Table 6: Comparing 2006 and 2011 spatial concentration of disadvantage in Sydney, Melbourne and Brisbane Sydney

Melbourne

Brisbane

All

No. of disadvantaged suburbs 2006

91

50

36

177

No. of disadvantaged suburbs 2011

100

49

38

187

Absolute change 2006–11

+9

-1

+2

+10

% of suburbs disadvantaged 2006

11

10

9

10

% of suburbs disadvantaged 2011

13

10

9

11

2

0

0

1

% of CDs/SA1s in disadvantaged suburbs 2006

72

77

61

72

% of CDs/SA1s in disadvantaged suburbs 2011

75

77

65

74

Absolute change 2006–11

+3

0

+4

+2

No. of disadvantaged suburbs 2006—decile threshold

38

18

12

68

No. of disadvantaged suburbs 2011—decile threshold

45

21

17

83

Absolute change 2006–11

+7

+3

+5

+15

Absolute change 2006–11

Source: Based on 2006 and 2011 Census analysis. Note: Except where specified otherwise, disadvantaged suburbs classified in relation to SEIFA quintile threshold.

33

In interpreting these findings it is, however, important to recognise the risk of incurring the so-called ‘ecological fallacy’ (Darcy & Gwyther 2012; Goldie et al. 2014). Our classification rates areas according to the local population’s socio-economic disadvantage in aggregate. Hence, not all of those living in a ‘disadvantaged area’ will necessarily be disadvantaged individuals. Similarly, many disadvantaged individuals living in the three cities will be resident in localities not classed as such.

32

Similar ‘poverty concentration’ processes have recently been observed in other countries. In the US, the 2000–11 period saw a renewed increase in the spatial concentration of low-income populations. While the number of ‘high poverty’ census tracts rose by 50 per cent, the concentration of poverty (the percentage of poor persons living in high poverty neighbourhoods) rose from 10.3 per cent to 12.8 per cent (Jargowsky 2013)34. Distinct from Australia, however, is that the spatial concentration of poor people in the US—albeit recently on the rise—has yet to surpass its 1990 peak value of 15.1 per cent (Jargowsky 2013). Similar ‘economic segregation’ trends have been observed in other countries comparable with Australia. Albeit measuring a slightly different phenomenon, Canadian research focused on Toronto in the period 1970–2005 found ‘a continuing trend towards the creation of a city with increasing disparities between rich and poor neighbourhoods’ (Hulchanski 2010, p17). As revealed by detailed statistical analysis, this process of divergence was leading to ongoing erosion in the representation of ‘middle income’ neighbourhoods. Similarly, another recent Canadian study found evidence to support the contention that ‘social differentiation’ has been generally on the increase in the nation’s urban centres and that, in most of the cities concerned, the two decades to 2006 had seen ‘a rise in the spatial concentration of low-income populations’ (Ades et al. 2012 p.355). Such changes have been attributed to labour market developments; in particular the growing polarisation of recent decades as earnings growth in financial and other services has substantially outstripped that in low-skilled occupations. Thus, it is argued that ‘polarisation of the labour market has resulted in a clear division of urban space’ (Ades et al. 2012, p.342).

3.3.3 The dynamics of socio-spatial change Beyond observations on changes over time in the overall quantum of disadvantaged suburbs, additional insights can be gained by analysing the finer-grained patterns of change in terms of individual suburbs becoming (or ceasing to be) disadvantaged. The ‘net change’ figures generated by our study (see Table 6 above) somewhat obscure the actual extent of change. Thus, while the net increase in disadvantaged areas was only 10, the gross number becoming disadvantaged in 2011 was 27. With three-quarters of these being contiguous with existing disadvantaged areas, we have another finding consistent with the spatially concentrating tendency demonstrated above. Classifying the 27 ‘emerging disadvantage’ suburbs according to the defining features of our typology categories is also revealing. Of the 18 such areas which could be unambiguously assigned to one of the four categories, eight were akin to Type 1 areas, while six were similar to Type 3 areas—in both cases a far greater representation than within the overall cohort of 2006 disadvantaged areas (see Table 12 in Hulse et al. 2014a). This is important given the Hulse et al. observation that areas of these kinds posed the most important challenges from a policy perspective. At the same time, among the 17 suburbs that ceased to be disadvantaged between 2006 and 2011, seven were Type 4 localities, a slightly disproportionate representation consistent with their housing market label as ‘dynamic improver’ areas (see Table 5 above).

3.3.4 The ongoing suburbanising dynamic: recent trends Table 7 below connects the above ‘dynamics of change’ analysis with the changing geography of socio-economic disadvantage. Albeit using a rather ‘broad brush’ measure, the key finding here is the observed ‘outward shift’ of socio-economic disadvantage away from central cities, suggesting continuation of a trend observed in relation to the period to 2001 (Randolph & Holloway 2005b). Thus, in Sydney in 2006 disadvantaged suburbs were, on average, located 33 kilometres from 34

Unfortunately no direct comparison can be made from published Australian census data—our own analysis is based on SEIFA values for CDs as geographical entities rather than relating to disadvantaged (or ‘poor’) individuals—see earlier footnote on the ‘ecological fallacy’. Another important factor to be borne in mind here is the potential influence of analytical scale. Averaging some 4000 inhabitants, a US census tract is more akin to an Australian suburb than a CD (United States Census Bureau 2010).

33

Sydney CBD whereas, by 2011, this had increased to 34 kilometres. Even over this short period the ‘epicentre of disadvantage’ moved outwards by at least 1 kilometre in all three cities. Moreover, again in all cases, neighbourhoods ceasing to be disadvantaged during this period were nearer to CBDs than the cohort-wide norm. Meanwhile, 2011 ‘emerging disadvantage’ localities were typically positioned ‘beyond’ previous epicentres. Table 7: Average distance from CBD to disadvantaged suburbs (kilometre) All disadvantaged suburbs 2006

All disadvantaged suburbs 2011

Only disadvantaged in 2006

Only disadvantaged in 2011

Sydney

33.0

34.2

28.4

37.5

Melbourne

21.9

23.6

15.5

35.2

Brisbane

27.1

28.2

21.9

30.5

All

28.7

30.2

23.1

35.6

Source: Based on 2006 and 2011 Census data

By comparison with some other developed countries, Australia’s now predominantly suburban location of concentrated disadvantage is pronounced. In the US ‘the poor [remain] disproportionately concentrated in central cities’ (Jargowsky 2013, p.22) and a similar observation could be made for the UK (Hunter 2014; Price 2015). In their recent spatial trajectories, however, the US, Canada and the UK appear to have been closely paralleling Australia’s experience. In the US, also incorporating a ‘de-clustering’ of high poverty census tracts, this has featured ‘a marked movement [of poor populations] … away from the downtown core …’ (Jargowsky 2013, p.13). Similar trends have been identified in Canada’s major cities; in the case of Toronto already emergent as long ago as the 1970s (Walks 2001). Based on their analysis of Montreal, Toronto and Calgary in the period 1986–2006, Ades et al. (2012) observe that ‘… the zones where poverty has increased in the last twenty years are located in areas further away from the downtown core than they were in 1986’ (p.357). In the UK meanwhile, the period 2001–11 saw a disproportionate increase in the number of suburban areas with above average levels of poverty. In London, for example, the 2001–11 analysis showed a 1 percentage point poverty increase in suburban areas compared with a 3 per cent fall in inner areas (Hunter 2014). More generally, an analysis focused on England’s eight largest metropolitan areas showed that the suburbanisation of poverty was occurring more sharply here than in smaller cities. Likewise, in Glasgow, it has recently been observed that poverty became ‘less centralised’ in the period 2001–11 (Price 2015).

3.3.5 The ongoing suburbanising dynamic: drivers of change In seeking to account for the observed ‘suburbanisation of disadvantage’ over recent decades a number of drivers have been suggested. Referring to the process as it affects Sydney, Randolph & Holloway (2005b) highlighted the possible role of selective migration whereby ‘upwardly mobile’ populations of ‘older’ (lower quality) suburbs depart to higher status areas, to be replaced by predominantly lower income successor populations, including recently-arrived migrants. Expanding on this explanation, Randolph and Holloway (2007) argued that an important contributor to Sydney’s increasingly suburban pattern of disadvantage has been the location and built form of the city’s middle suburbs, areas primarily built in the three decades to 1970. Latterly, these areas had experienced the problems of a cheaply built and deteriorating housing stock which, because of its relatively low value, had proven attractive to poorer households excluded from both inner and outer suburban areas. The age of housing and its depreciation have also been cited as possible explanations for poverty suburbanisation in the Canadian context. Thus, older suburbs where housing stock is subject to lack of investment may be selectively drawing in disadvantaged populations (Ades et al. 2012, 34

p.357). In the specific instance of Toronto, suburbanisation of deprivation in the late twentieth century was attributed to the decentralisation of social housing (Ley & Smith 2000). In the US and UK contexts, two contributory factors have been recently cited. First, existing suburban populations have become poorer. And, second, suburban areas have increasingly accommodated lower income populations excluded from inner cities by rising housing costs resulting from ‘revived local housing markets’ (Berube 2014; Hunter 2014). As implicit in this latter observation, the underlying dynamic in the suburbanisation of disadvantage would seem to be a steepening urban land value gradient consistent with the ‘revalorisation’ of the inner city (Atkinson & Bridge 2004). Explanations emphasising the filtering of well-located older housing into use by poor people are perhaps being consequentially rendered less significant. In explaining the underlying causes of growing US suburban poverty, Berube (2014) notes that this partly reflects the fact that in some metropolitan regions new low-skilled immigrants settle predominantly in suburbs, rather than in inner city areas. Additionally, the ‘Great Recession’ hit US suburban job markets particularly hard—for example as regards construction and manufacturing, two major industries particularly concentrated in outer suburban locations. While the first of these two factors is undoubtedly echoed in Australia, it is less clear that the second is directly paralleled in part because Australia has had continuous economic growth since 1994. Rather, the longer-run restructuring of the Australian economy away from manufacturing and towards services has been associated with ‘a centripetal force pulling economic activity and housing investment back into the centr[al city]’ (Randolph 2015). We return to this issue below. As regards ‘housing policy influences’, the post-1990 suburbanising and de-clustering US trends depicted by Jargowsky (see above) could possibly have been influenced by the HOPE VI public housing redevelopment program. Partly justified by a stated aspiration to ‘de-concentrate poverty’, this had by 2010 seen the demolition of over 150 000 public housing dwellings, with only just over half this number being ‘re-provided’ within the context of mixed tenure (and/or mixed use) redevelopment of cleared sites (Schwartz 2015). What makes this potentially relevant in the current discussion is the inner urban location of many formerly ‘distressed estates’ covered by HOPE VI, and the resulting dispersal of many former residents. However, while this may have contributed to the wider ‘poverty suburbanisation’ trend in certain cities, the program is considered much too small to have constituted a ‘main driver’ of such trends across urban America as a whole (Galster 2015). Although its national influence in this respect is undoubtedly limited, it could be argued that the HOPE VI contribution to the US suburbanisation of disadvantage is a ‘policy-driven’ contribution to the suburbanisation of disadvantage. There has been no directly equivalent factor in the Australian context.35 However, to the extent that the process results from the increasing ‘housing market exclusion’ of low-income groups from inner city areas (see above), it is arguably more relevant to highlight the absence of policy—in this case the lack of any practical strategy to protect and expand a stock of inner city affordable housing (i.e. accommodation ‘affordable’ by low-tomiddle income households). In cities such as New York and London, such policies do operate, at least to an extent. Whether through control of rents during tenancy or inclusionary zoning obligations on housing developers, state bodies play an active ‘market-mitigating’ role in preserving and expanding affordable housing. In Australia, however, especially given their comprehensive appraisal in AHURI reports and elsewhere, there can be no doubt that governments have made active choices to forgo such interventions.

35

While acknowledging the recent NSW Government initiative to privatise public housing at Millers Point, Sydney, this at present remains a relatively small scale and isolated initiative.

35

3.3.6 Housing market processes and spatial disadvantage: evidence from this study Social housing Linked with the above discussion, we now consider the wider issue of how housing transactions, preferences and trade-offs shape the socio-spatial structure of cities; in particular, how they contribute to the spatial concentration of disadvantage. The fundamental ‘sifting and sorting’ role of housing market processes here is long-recognised (Hamnett & Randolph 1988). The abovecited Randolph and Holloway comments tap into this thinking. Given the Australian tradition of equating urban disadvantage with public housing, associated commentary has often focused on the impacts of the increasingly ‘targeted’ approach to housing allocations seen over recent decades (Jacobs et al. 2011). The growing preference accorded to ‘greatest need’ applicants is widely recognised as having compounded the ‘residualisation’ of the social rental tenure (i.e. public and community housing). Certainly, many case study area interviewees involved in our research emphasised the links between the disadvantaged status of their suburb and housing system factors as mainly involving the local scale of public housing and associated managerial policies (see Cheshire et al. 2014). However, as noted above (Section 3.2) public housing accommodates only a small proportion of disadvantaged suburb populations. Moreover, the extent to which associated allocations processes contribute to the spatial concentration of disadvantage also needs to be unpacked. The hypothesis here posits a form of selective migration in which the presence of social housing in a locality funnels highly disadvantaged people into the area from more advantaged places, replacing ‘aspirational’ former tenants whose improved economic situation enables them to ‘transition up’ to private housing in a higher status area. Albeit based on only limited evidence, our study provides little backing for this theory. This was explored in the context of our Emerton/Mount Druitt (Sydney) fieldwork—the area containing the largest proportion of public housing in any of our six case study localities (26% in 2011). Housing Department records showed that most of those being allocated public housing tenancies here were people already local to the area (e.g. being rehoused from unaffordable or otherwise unsatisfactory private rental). Almost three-quarters (74%) of the 203 public housing lettings in 2013 involved new tenants who already had a Mount Druitt address. The main spatial impact of public housing allocations policies, therefore, was to concentrate disadvantaged people from within the locality rather than to funnel disadvantaged people into the area from elsewhere in Sydney. At the same time, tenancy turnover in the area was relatively low, with newly arising vacancies generally resulting from deaths, local transfers, incarceration or rent arrears evictions. In only a very small proportion of instances could newly arising vacancies be attributed to ‘aspirational moves’ where a former tenant, having improved their financial circumstances, exits to a ‘better area’. In part, this latter finding reflects the overall impoverishment and consequential ‘silting up’ of Australia’s social housing due to lack of affordability for lower income households of the main options: home ownership and private rental housing. Bearing the above in mind, it would be something of a caricature to portray the Mount Druitt public housing vacancy-generation and letting system as predominantly involving socially mobile outmovers replaced by highly disadvantaged in-movers. There seems little reason to believe that these findings are atypical of social housing in disadvantaged areas. Private rental housing Across our cohort of disadvantaged suburbs in Sydney, Melbourne and Brisbane, the incidence of private rental housing, at 27 per cent, was almost three times as great as social rental—10 per cent. Significantly, by comparison with city-wide norms, disadvantaged suburbs had seen disproportionate growth in private rental housing over the 2001–11 period (Hulse et al. 2014a; Cheshire et al. 2014).

36

Moreover, in areas of this kind, the private rental sector is home to a predominantly lower income population. Some 29 per cent of private renters in our residents' survey reported monthly household incomes of under $200036 and 78 per cent received less than $5000—an even higher incidence than that for public housing (70%).37 (Pawson & Herath 2015). And, albeit slightly lower than the comparable figure for public housing (40%), some 36 per cent of private renters were classed as ‘economically excluded’, taking account of the extent to exposure to specific forms of ‘income deprivation’ during the previous year. 38 Factoring in their much greater numerical representation in disadvantaged places, private renters were estimated as accounting for 56 per cent of all economically excluded residents of such areas in Sydney—as compared with the 29 per cent who were social renters (Pawson & Herath 2015). Also as revealed by our survey, a quarter of the entire private rental population of the fieldwork areas had arrived from other areas within the previous five years. While further research would be required to substantiate this, it is possible that many such in-movers will have been drawn into their new area from less disadvantaged places by the magnet effect of relatively affordable housing in a very low status area. If the availability of private rental housing in a disadvantaged area is associated with the attraction of low-income households, and such areas have been recently subject to the disproportionate expansion of such housing, questions are raised about the factors underlying this dynamic. Fieldwork evidence from our case study work suggests that relevant (albeit partially overlapping) factors may include:  Disproportionate recent investor landlord activity in purchasing new and existing housing lower

value areas.  Active disposal of public housing in certain localities and direct conversion to private rental.  Second home owners being reluctant to sell in weak housing markets.

The associated evidence is recounted in some depth in Cheshire et al. (2014). Home ownership While rental housing was overrepresented in disadvantaged suburbs, owner occupied homes still accounted for a majority of homes (57%) in the identified areas. Owners were split almost evenly between ‘owned outright’ and ‘buying with a mortgage’, with survey evidence suggesting that most of the former (81%) were confined to lower incomes—that is less than $5000 per month (Pawson & Herath 2015). Collectively, however, owner occupiers living in disadvantaged areas were estimated as accounting for only 15 per cent of the ‘economically excluded’ population of such areas (Pawson & Herath 2015, Figure 15).39 Residents' survey evidence about owner occupiers also contributes to our exploration of the links between the operation of housing markets and the dynamics of spatial disadvantage. This suggested that disadvantaged suburbs play an important role as ‘home ownership gateway’ areas for ‘aspirational’ households intending subsequent moves to other (probably higher status) areas. Thus, two-thirds of those recently purchasing homes in disadvantaged areas for their own 36

To contextualise this figure ($USD1525) readers should note that the median gross monthly household income for Greater Sydney recorded in the 2011 Census was $6222 (see Pawson & Herath 2015, Table 9). 37 Discounting the 25 per cent of survey respondents who declined to answer the income question. 38 Here we adopted Bray’s suite of ‘hardship measures’ originally developed by the ABS (Bray 2001) and still used in ABS surveys. Here, survey respondents are asked whether a lack of income had meant that they or fellow household members had, within the past 12 months, needed to forgo certain essentials (e.g. winter heating or meals), had needed to pawn possessions or had needed to seek assistance from family or welfare organisations. 39 Relevant to this finding, 77 per cent of outright home owners were aged over 60 and 42 per cent were aged over 70. These statistics are based on the age of the survey respondent. However, among owners buying with a mortgage the age profile was very different. Here, 42 per cent of respondents were aged under 40, with 71 per cent under 50. Overall, therefore, home owners had a fairly even age profile—26 per cent of all owner respondents were aged under 40, while 48 per cent were aged over 60.

37

occupancy (70%) had moved to the area from ‘beyond the locality’—a far higher proportion than the comparable figure for newly resident renters in such areas (43%). It seems likely that, for at least some of the ‘in-mover homeowners’, this will have been a transition from a higher status area in which house purchase would have been unaffordable. This accords with the housing market analysis which showed that house prices in disadvantaged suburbs represent a substantial discount to city-wide medians unlike rents which are relatively close to city-wide medians (Hulse et al. 2014a, pp.44–50). Likewise, among disadvantaged area owner occupiers buying with a mortgage, almost half (49%) agreed with the statement ‘I would get out of this area if I could’, a figure substantially higher than for the other tenure groups (28–39%). Also, among all owner occupiers aspiring to make a housemove, a substantial majority (63%) anticipated that their destination would be beyond the locality. Again, this was well above the norm for other tenure groups (54% for public renters and 46% for private renters). These findings appear consistent with hypotheses around selective mobility as a significant contributor to socio-spatial urban patterns, whereby more aspirational households use lower status areas as a stepping stone in housing careers and with the detailed housing market analysis conducted for this study. The data cited above are presented and explained in more detail in Pawson & Herath (2015).

3.4

Places that disadvantage people

3.4.1 The ‘neighbourhood effects’ debate The term ‘neighbourhood effects’ refers to the idea that living in a poor neighbourhood can compound the impact of poverty and disadvantage on an affected individual. In other words, ‘deprived people who live in deprived areas may have their life chances reduced compared to their counterparts in more socially mixed neighbourhoods … living in a neighbourhood which is predominantly poor is itself a source of disadvantage’ (Atkinson & Kintrea 2001, pp.3–4). Expressed in more economic terms, neighbourhood effects are ‘the consequences of spillovers and externalities that arise from the co-location or proximities of particular socio-economic groups or activities’ (Maclennan 2013, p.271). In his 2012 meta review of neighbourhood effects research, Galster identified 15 distinct mechanisms through which such effects may be generated, split into four broad groups as shown in Table 8 below. Conclusions from Galster’s review of research evidence associated with these different types of ‘neighbourhood effect’ are summarised in an earlier report in the current series (Pawson et al. 2012).

38

Table 8: Neighbourhood effects typology Broad rubrics

Specific mechanisms

Social-interactive mechanisms

Social contagion Collective socialisation Social networks Social cohesion and control Competition (for resources/opportunities) Relative deprivation Parental mediation

Environmental mechanisms

Exposure to violence Physical surroundings Toxic exposure

Geographical mechanisms

Spatial mismatch/accessibility Public services

Institutional mechanisms

Stigmatisation Local institutional resources Local market actors

Source: Derived from Galster 2012

As noted in Section 1.2.5, the neighbourhood effects hypothesis has evoked extensive policymaker interest. Demonstrating official recognition of the notion in the Australian context, a 2009 Commonwealth Government-commissioned report commented: ‘It has been found that when social disadvantage becomes entrenched within a limited number of localities a disabling social climate can develop that is more than the sum of individual and household disadvantages and the prospect is of increased disadvantage being passed from one generation to the next’ (Vinson 2009, p.2). In the academic world, the thesis has sparked lively contestation, especially in association with US government initiatives to ‘de-concentrate poverty’. For contributors such as Slater (2013), the concept is inherently suspect because of its supposed origins in debates involving Murray’s ‘underclass’ thesis. As summed up by Arthurson (2013), ‘The nub of the [underclass] debate was the extent to which individuals who live in poverty are culpable for their own predicament (individual agency) and the degree to which societal (structural) factors affect individual capabilities, in combating or adding to disadvantage’ (p.254). Another reason for the prominence of these deliberations has been the contention that if ‘neighbourhood effects’ do not in fact exist then there is no justification for place-based urban policy at all. Manley et al. (2013) argue that if neighbourhood effects do not exist on any scale, then ‘neighbourhood intervention is merely redistributing resources or opportunities to residents there at the expense of groups outside the neighbourhood—a zero sum game’ (p.5). Notwithstanding this point, Maclennan (2013) contends that (at least in Europe) government commitments to neighbourhood renewal programs have stemmed not from ‘a belief in evidence of ‘neighbourhood effects’’ but, rather, from other aspects of the new public management, ‘in particular, the importance of integrated and preventative approaches in public services and growing roles for community voice in provision’ (p.270). According to one respected expert’s recent contribution, ‘a consensus seems to have emerged among social scientists that neighbourhoods do indeed matter in determining human welfare across a variety of salient dimensions … a growing number of studies have used a range of methodologies to demonstrate the negative effects of concentrated neighbourhood disadvantage on human well-being, especially over the long term’ (Massey 2013, p.692). However, also

39

attempting a resume of current thinking, Tunstall (2013) contends that ‘most [urban studies] researchers appear to be generally sceptical that neighbourhood effects alone provide a justification for costly or disruptive neighbourhood policies’, and that this reflects ‘a combination of the absence of good evidence and, increasingly, good evidence of weak effects’ (p.178). As it has influenced policy thinking in Australia, Arthurson (2013) suggests that the notion of neighbourhood effects underlies a policy-maker perspective in which the spatial concentration of social housing is seen as ‘a threat to social cohesion’, a line of thinking that gives rise to the estate renewal model in which the creation of ‘social mix’ is a central objective. However, while this thinking appears influenced by US debates, ‘the different social and political context in Australia makes its application less relevant in this milieu’ (Arthurson 2013, p.253). Meanwhile, in Arthurson’s view, the adoption of the ‘social exclusion’ terminology in Australia is motivated by a wish to ‘maintain distance from US debates’ and their ‘perjorative’ undertones.

3.4.2 Impacts of spatially concentrated disadvantage: evidence from this study Especially given the need for counter-factual scenarios, rigorously researching the existence and scale of neighbourhood effects is a highly demanding enterprise, well beyond the capacity of the current study. Nevertheless, empirical data collected through the study—both through the residents' survey and the locality case studies—have some bearing on the issue, especially in terms of Galster’s ‘geographical’ and ‘institutional’ mechanisms (see Table 8 above). Before discussing these, however, we first draw on residents' survey data to gauge the extent to which our ‘disadvantage concentrations’ fully merit that label, and then for an overview of ‘disadvantaged area’ resident views on their locality. Depth of spatial disadvantage Neighbourhood effects are arguable only where income poor or otherwise socially excluded populations are spatially concentrated to a substantial extent. While the areas we have identified as ‘disadvantaged suburbs’ rate as such in relative terms, what is the depth of that disadvantage? One Canadian study addressing this issue at an international level found that levels of spatial polarisation in metropolitan Canada exceeded those of counterpart Australian cities. In turn, ‘degrees of segregation’ in Canadian cities were ‘lower than many cities in the United States and Britain’ (Walks & Bourne 2006, p.294). Similarly, collated as part of the current study and drawing on existing literature, Burke & Hulse (2015) concluded that ‘the depth of disadvantage [in urban Australia] is nowhere as great as in the deindustrialised areas of many equivalent international cities’ (p.10). Related to this question is the potential for a study of the current type to incur the ‘ecological fallacy’—that is ‘an error of deduction that involves deriving conclusions about individuals solely on the basis of an analysis of group data’ (O’Dowd 2003, p.84). This arises from analytical reliance on census-derived statistics for spatial units (see Section 3.2). Given their status as aggregate measures, these could potentially obscure or mislead in terms of the characteristics of the individual persons and households within each spatial unit concerned. While high in aggregate terms, how does the actual incidence of disadvantage within our ‘disadvantaged areas’ rate alongside appropriate comparator numbers? As demonstrated by the residents' survey (see Table 2 and Section 1.4.5), the occurrence of significant economic hardship in our ‘disadvantaged areas’, while far from universal, was substantially above national norms. In terms of recently having had to forgo necessities, having experienced problems in paying for essential items or services or in having had to seek external financial help, an average of 33 per cent of households in the four survey areas had been directly affected by financial poverty during the previous year, two-thirds higher than the national (and Sydney-wide) norm (20%). Such deprivation rates were, thus, typically 65 per cent ‘above normal’. While deprivation rates were highest among public renters (at 50%), the rate for private tenants (41%) only fairly marginally lower. Homeowners—especially outright owners—were far less affected. 40

Especially in the light of the fairly extreme tenure-specific contrasts, we can also reason that our findings are highly influenced by our chosen analytical scale. As noted in Section 1.4.2, utilisation of ABS suburbs as our prime areal unit (population typically 4–8000) was based on a reasoned judgment. However, if instead we had used Collection Districts (average population approximately 600) there would have been many more identified areal units dominated by a single housing tenure—including a significant number containing only public housing. This would have impacted on our conclusions as regards depth of disadvantage in specific localities. Relevant to policymaker concerns, these observations also highlight the emptiness of discussions on ‘mixed communities’ where ‘acceptable threshold’ levels of public housing within a locality are cited without any firm grounding in stated assumptions about analytical spatial scale. Residents' views on local neighbourhoods While revealing relatively high deprivation rates, however, residents' survey results at the same time confound any suggestion that Australia’s disadvantaged suburbs are ‘sink areas’ evoking little local loyalty or pride. As shown in Table 9 below, two-thirds of residents (68%) expressed a feeling of local belonging. Similarly, as fully reported elsewhere (Pawson & Herath 2015), the balance of opinion in such localities was that local areas had recently been experiencing positive change. Nearly a third (32%) took this view, as compared with only just over a fifth (22%) who perceived recent decline. Moreover, well over half (62%) believed their area to have a strong sense of community, with almost half (49%) reporting membership of a local community group or club (usually a social or sports club). Despite the unfortunate dearth of city-wide or national comparator numbers, these statistics seem consistent with the Burke & Hulse (2015) conclusion that, at a neighbourhood level, poverty in Australia is relatively ‘shallow’. At the same time, however, other findings on residents' views about their localities appear consistent with the expectation that such areas are, to some extent, ‘troubled’. While again acknowledging that the lack of external benchmarks makes these statistics somewhat difficult to interpret, the incidence of concerns about neighbourhood social problems such as car hooning, 40 crime and drug abuse appear high (see Table 9). However, it is not clear how such perceptions relate to more objective data on the incidence of crime and antisocial behaviour. Table 9: Respondents' views on their locality by housing tenure Statement

Owner

Purchaser

Private renter

Public renter

All tenures

% of respondents agreeing or strongly agreeing with statement My local area is a safe place to live

79

63

68

60

69

I feel I belong in this neighbourhood

73

49

71

69

68

Physical appearance is appealing

33

44

59

59

50

Car hooning is a problem here

70

40

49

49

53

Crime is a problem here

45

57

38

49

41

Drugs are a problem here

36

57

35

49

40

Graffiti and vandalism are problems

34

36

28

42

33

Nuisance behaviour from excess drinking is a problem

10

32

33

36

27

I would get out of this neighbourhood if I could

28

49

39

37

37

146

93

256

250

745

N (minimum) = Source: reproduced from Pawson and Herath 2015 40

Misbehaviour involving reckless driving.

41

Geographical ‘neighbourhood effects’ mechanisms: public service provision Observations that concentrations of poor people tend to coexist with poor quality public services are longstanding (Powell et al. 2001; Fisher & Bramley 2006). Such contentions have a particular logic in the United States where a highly decentralised service provision framework for services such as justice and education, combines with the substantially self-funded status of municipalities. This means that a weak or decaying local tax base associated with an impoverished population is liable to undermine important contributors to local quality of life. However, concerns that the residents of disadvantaged places suffer sub-standard public services have also been widely cited in the UK, despite the equalisation formula built into the country’s local government finance system. A recent UK study, for example, found clear evidence of ‘environmental injustice’ in the form of poorer street cleaning services in less affluent areas. ‘[Street cleaning] is supposed to be a universal public good, yet outcomes are significantly worse for deprived groups and areas. Their social and economic disadvantages are compounded by having to experience dirtier, less attractive streets and public spaces’ (Bramley et al. 2012, p.758). By comparison with their American counterparts, it has been argued that poor localities in Australia are protected by key features of the urban governance framework—in particular, the state (rather than municipal) provision of key public services such as education and justice (Burke & Hulse 2015). Nevertheless, Australian research suggests that schools drawing on areas with disadvantaged populations will generally record lower achievement than national norms. This follows from the empirical finding that ‘school average student characteristics (particularly socioeconomic indicators) are very strong predictors of school average performance’ (Holmes-Smith 2006, p.2). While it has no direct implication for the quality of the educational service (i.e. as in a ‘value added’ measure), this observation is consistent with the idea that schools in ‘disadvantaged places’ are likely to be ‘low performing’ establishments—and reputed as such.41 In some of our case study areas interviewees commented negatively about the run-down appearance of local schools: ‘You look at some of the schools around here and I’m sure you could find prisons that look better’ (public housing official). However, while there was some implicit acknowledgement that problematic school quality might have disadvantaged local people in the recent past, some parent groups remarked on what they saw as ongoing improvements. Backing this up, census evidence showed reduced rates of early school leaving for all case study areas, with some out-performing their respective metropolitan area. A more common case study area concern, as regards educational provision, was the perception that the local structure of secondary education disadvantaged certain pupils by forcing them to travel to distant schools. In one area—Russell Island, Brisbane—this reflected locational remoteness and the Queensland Government view that people choosing to reside on the island did so in the knowledge that such services were not locally provided. In another area—Auburn, Sydney—the issue arose more from educational gender segregation; with the only public secondary school being for girls, boys were necessarily obliged to out-migrate daily from the age of 11. Provision of health, support and social service facilities accessible to disadvantaged places evokes a slightly more two-edged debate. There is an argument that the sparcity of such provision in disadvantaged communities could be another ‘disadvantaging’ factor negatively impacting on local residents. Some research interviewees were, on the other hand, concerned that where such services were concentrated close to the population they served, this could institutionalise an area’s disadvantaged social profile. Indeed, in one case study area it was contended that the locally high density of social services might only compound the area’s negative external image (see below). Similarly, other participants (including community workers and housing managers) saw a risk that easy local access to such provision might even promote ‘welfare dependency’. In 41

It must be acknowledged that the relatively large scale of private education in Australia (a third of students at nongovernment schools) complicates analysis of this issue.

42

any event, based on interviewee accounts, the six case study localities spanned something of a continuum from service rich to service poor locations. In any event, residents' survey results reveal scant evidence of widespread local dissatisfaction as regards accessibility of schools or health services in disadvantaged areas—see Table 10 below. While the survey fieldwork areas were not strictly representative of all disadvantaged suburbs (see Section 1.4.5), these figures strongly suggest that relatively few residents of such areas perceive themselves to be locationally underprivileged in this respect. The stark contrast with perceptions on the accessibility of employment-rich locations demonstrated in Table 10 is a finding to be revisited in Section 3.5. Table 10: Disadvantaged area accessibility to services and employment: residents’ perceptions Statement

% of respondents in agreement with statement

The area is well-served by public transport

84

The area has good access to primary schools

90

The area has good access to secondary schools

76

The area has good access to health services

84

There are good employment opportunities within or accessible to the area

32

Source: Residents' survey. Note: Percentage figures are an average of respective scores across each of the four survey localities.

Institutional ‘neighbourhood effects’ mechanisms: area stigmatisation As elicited in the locality studies, resident views about living in an area of concentrated disadvantage mainly focused on the issue of negative stigma. In Galster’s (2012) classification of NE processes, this was classed as one of three identified ‘institutional mechanisms’ (see Table 8). Living in a place with a negative reputation can compound the experience of disadvantage, affecting an individual’s health and wellbeing (Bauder 2002; Permentier et al. 2007; Kelaher et al. 2010). Consistent with conclusions from numerous existing studies (e.g. Hastings & Dean 2003; Kearns et al. 2013), there was a keen awareness of the ‘area stigmatisation’ issue among stakeholder and resident interviewees in most of our case study localities. Usually considered as grossly distorted or outdated, these unflattering images were often seen as originating in, or perpetuated by, the media. In some cases, such press coverage was considered a gratuitous reinforcement of negative images. Based on his extensive early 1990s fieldwork in four poor urban neighbourhoods across Australia, Mark Peel argued that community life in such areas had already fallen victim to a lazy journalistic need for locality stigmatisation as ‘the chief illustration for stories about a looming social crisis’ (Peel 2003, p.16). Among instances relating to Mount Druitt, Sydney, and illustrating contemporary concerns around this issue was the TV news treatment of a 2013 visit to Western Sydney by then Prime Minister Julia Gillard when, as highlighted by a case study interviewee: … the only people they interviewed in [the Mount Druitt] mall were people who conformed to the stereotype (tattooed, wearing thongs, appearing drunk)—that’s the story they wanted to tell. [Local people] … do find it really distressing. (NGO community worker/support provider) Area stigmatisation stemming from, or at the very least, recycled through media representation is evident in the following press extracts:

43

Not long after Athens declared itself the cultural capital of the known world it was over-run by uncouth barbarians, namely the Macedonians, who then went on to conquer half of Asia. By way of perspective, it was a bit like if Woollahra had been taken over by the residents of Mt Druitt, who then went on to invade New Zealand. ('Laziness the key to Greek plan for world domination', Daily Telegraph 25 August 2012) Charlotte Feldman, a member of the [Darling Point] action group, dismissed suggestions anyone had harassed Mrs Jones. 'This is not Mount Druitt. People know how to behave'. ('Locals in a lather over Moran heir's shindigs', Sydney Morning Herald 25 February 2012) Also revealed by our systematic media analysis was the following more reflective journalistic acknowledgement that: Part of our fascination with the story of a missing Mount Druitt schoolgirl, Kiesha Abrahams, is that it speaks to and reinforces various stereotypes we have about class and parental neglect. ('Harrowing film opens our eyes to child abuse', Sydney Morning Herald 8 September 2010) As argued by some case study interviewees, this was more than a matter of hurt local pride, since residence in an affected area was claimed to negatively impact on residents' experiencing discrimination by potential employers. Whether such beliefs could be objectively evidenced remains an open question beyond the scope of the research. What is, however, clear is that that some residents strongly believe this to be the case. If there were substance in this it could perhaps be considered a weak instance of ‘neighbourhood effects’.

3.5

Impacts of the observed geography of concentrated disadvantage

Distinct from the question of whether the spatial grouping of deprived people compounds the problems experienced by individuals through neighbourhood effects ‘local to the place’ are the impacts of the broader geography of disadvantage. As shown in Sections 3.2 and 3.3, concentrations of disadvantage in Australia’s largest cities were, by 2006, located largely in middle and outer suburbs and were continuing to move outwards away from CBDs in the 2006–11 period. This needs to be seen within the context of the economic restructuring ongoing in Australia over the past 20–30 years, involving the gradual replacement of manufacturing (substantially decentralised) by finance and services (more spatially clustered in CBDs and—to a limited extent—secondary urban nuclei). Emphasising the extent of this change Rawnsley & Spiller (2012) reflect that ‘ … the economic base of Melbourne at the start of the second decade of the 21st century is barely recognisable compared with the industries which underpinned the city in the 1980s’ (p.138). Thus, Australia’s major cities have become ‘exporters of ideas, technology, connections and experience, as distinct from manufactured goods’ (Rawnsley & Spiller 2012). Given these trends, residents of middle and outer suburbs are increasingly excluded from the associated ‘higher order’ job opportunities concentrating in the metropolitan centre. Consistent with this observation is the above-reported research finding that less than a third of disadvantaged area residents in Sydney considered their area to be within range of good employment opportunities (see Table 10). Indeed, fully 50 per cent of residents' survey respondents actively disagreed with this proposition.42 In Sydney, the ongoing economic restructuring of urban space is manifest in the way that ‘knowledge industries’ have been increasingly clustering within the so-called ‘global arc’ running from Macquarie Park through the CBD and south to Sydney Airport. Moreover, with the centre of gravity of Sydney’s disadvantaged population being pushed ever further South and West—away from this area—it is not only a matter of remoteness from where ‘higher order’ jobs are 42

That is, they disagreed or strongly disagreed with the statement: ‘There are good employment opportunities within or accessible to the area’.

44

concentrating. More broadly, as illustrated by figures from SGS Economics, Sydney’s entire economy has been tilting towards the global arc, with annual employment growth in this area averaging 2.1 per cent in the period 2008–13, as compared with only 0.5 per cent in Western Sydney (Wade 2014; Wade & Cormack 2014). Meanwhile, empirical research focused on Melbourne suggests that ‘workers in suburbs with superior connectivity … enjoy significant advantage in lifetime earnings. This reflects an advantage on behalf of the firms that employ them’ (Rawnsley & Spiller 2012, p.144). Such developments mean that growing remoteness from employment nodes is liable to increasingly disadvantage outer suburban residents. As observed by Kelly & Mares (2013), one important manifestation of this pattern is a huge suburban female economic participation deficit. Thus, in outer Melbourne, 2011 female participation rates are 20 per cent below male. This has been explained by Pocock et al. (2012) as connected with the ‘spatial leash’—the extent to which childcare responsibilities mean that mothers are constrained from working distant from their home localities. For individuals the result is unfulfilled potential, for society the impact is suppressed city productivity. Applying similar logic, Rawnsley & Spiller (2012) argue that Australia’s strategic planning and housing policies are detracting from improved urban productivity ‘particularly by placing the lion’s share of population growth in far-flung suburbs which, in some contrast to their counterparts from the early post-war period, now feature significantly inferior access to jobs and services’ (p.135). Thus, as well as being an equity issue, this is also ‘a major drag on productivity’ (Rawnsley & Spiller 2012). As emphasised by Burke & Stone (2014), for lower income households of Sydney and Melbourne, the exclusionary impacts of residence increasingly restricted to remote suburbs have been compounded by urban transport planning policies: ‘The remaking of labour market and housing markets in the last three of four decades, combined with strategic planning to accommodate the growth of the motor vehicle and disinterest in new public transport infrastructure investment or service patterns, has forced many lower-income households to confront new forms of financial hardship and transport disadvantage’ (p.17).

3.6

Chapter summary

While spatial concentrations of social disadvantage are clearly identifiable in Australia’s major cities, these are significantly diverse in terms of their socio-economic and housing market characteristics. By 2006, such areas were clustered primarily in the middle and outer suburbs of Sydney, Melbourne and Brisbane. Processes of spatial polarisation and suburbanisation of disadvantage have subsequently remained ongoing. Both of the above urban spatial dynamics are also common to other Anglophone developed nations such as the UK and US. However, Australia’s distinct social and urban heritage means that there are also important contrasts with these two comparator countries. While apparently continuing to escalate, the spatial concentration of poverty remains less intense in Australia— especially by comparison with the US. Thus, Australian ‘poverty neighbourhoods’ may be ‘shallower’ than in these other countries, yet the greater fluidity of Australian urban systems is evident in that geographically focused disadvantage has moved more decisively from inner urban localities to middle and outer ring suburbs than in comparator British or American cities. At least in areas towards the lower end of the ‘disadvantage ranking’ there are some local perceptions that local people are subject to a negative ‘neighbourhood effect’ but almost entirely in the form of stigmatisation of their suburb. However, while dissatisfaction with the locally accessible provision of public services such as health and education is only a minority view, the picture is different when it comes to employment. Consistent with the renewed dynamic towards a centralised, rather than dispersed pattern of economic activity in Australia’s major cities, residents of (primarily suburban) disadvantaged places are much more inclined to perceive their location as

45

problematic in this respect. The intensifying exclusion of lower income households from inner cities is compounding an urban social geography to the detriment of city productivity as well as individual welfare for members of this group.

46

4

ADDRESSING CONCENTRATIONS OF DISADVANTAGE: POLICY RESPONSES

4.1

Chapter overview

This chapter addresses the last of the three core research questions that have guided the study: how can policy, practitioners and communities respond to spatial disadvantage in ‘best for people, best for place’ terms? In doing so, the chapter draws primarily on the data generated through the in-depth local studies of selected disadvantaged areas to examine the kinds of interventions that have typically been implemented in the cities of Sydney, Melbourne and Brisbane to address spatial disadvantage and spatial concentrations of disadvantaged people. What has become clear throughout this program of research, however, is that there is considerable variation within the cohort of ‘disadvantaged areas’ in Australian cities according to: i) their distance from, and connection with, the metropolitan centre and other regional hubs; ii) the economic base within each area; iii) the composition of the local population, particularly in terms of cultural and ethnic mix; and iv) the local and metropolitan housing and labour markets, histories and dynamics. In other words, the characteristics and lived experiences of geographically concentrated socioeconomic disadvantage are highly spatially context specific. How responsive policy interventions are, or should be, to these local variations is an important question that this chapter seeks to address. It does this in two ways. The first is through an empirical assessment of recent policy responses to disadvantage in the six local areas of Auburn and Emerton/Mount Druitt (Sydney); Springvale and Braybrook (Melbourne); and Logan Central and Russell Island (Brisbane). The purpose of this assessment is not to provide an exhaustive list of the various policies, programs and projects that have been implemented in each site. Nor is it possible to undertake any kind of systematic evaluation of the interventions that have been put in place (Cheshire et al. 2014 provide a more detailed assessment of the perceived effectiveness of some of these programs from the point of view of local stakeholders). Rather, the objective is to detail the kinds of interventions typically targeted at disadvantaged people and places in Australia; to identify points of variation across different locality types; and to consider whether any observed patterns can be explained by the typology of disadvantaged areas outlined earlier in this report, which might suggest a spatially sensitive policy approach. The interventions examined here have typically taken one of two forms. The first are place-based solutions, which involve attempts to address problems associated with living in places of disadvantage (e.g. poor housing quality, limited transport access or high crime). The second are what Randolph (2004a) has termed ‘place-focused’ initiatives which target the problems facing disadvantaged groups within specific areas (e.g. a lack of job-readiness or low educational achievement). The main point of difference between these two initiatives is that where the former explicitly acknowledges the disadvantaging features of place, the latter adopts a more peopleoriented approach by seeking to improve the opportunities and outcomes of disadvantaged populations. In contrast to mainstream social welfare programs, however, Randolph contends that placefocused policies continue to be spatially relevant and actually operate in place for people ‘due to the fact that much of the activity they fund or support takes places in areas of high disadvantage’ (2004a p.65; see also Griggs et al. 2008). In this sense, they share an orientation to place that is absent from more mainstream social welfare programs. In mapping the application of these two basic types of interventions across the six disadvantaged areas, it becomes possible to identify variations and gaps in the way policy interventions have been targeted at, or otherwise impact upon, disadvantaged areas. These differences arise from the way certain types of localities appear to have missed out on large-scale funding programs and in the apparent failure of many

47

such interventions to properly target the fundamental problems that are common to disadvantaged localities regardless of the specific challenges they individually face. The second part of the chapter embeds these findings within international debates about the relative merits of place-based approaches to tackling disadvantage and their ability to be simultaneously ‘best for place’ and ‘best for people’. Once there seemed to be agreement that place-oriented policies offered a potential solution to the disadvantaging (i.e. detrimental) effects of poor neighbourhoods and the poor trajectories of their resident populations (van Gent et al. 2009). However, the logic of spatially-targeted strategies has become much more contested in recent years (see, e.g. Andersson & Musterd 2005; North & Syrett 2008; Maclennan 2013; Manley et al. 2013). Amid concerns over the inequitable distribution of resources across neighbourhoods, the ‘scalar mismatch’ between the causes of disadvantage and their policy responses (North & Syrett 2008; Rae 2011) and a lack of clarity over aims and outcomes (Manley et al. 2013), there is now doubt about the ability of spatially targeted interventions to address the wider systemic problems associated with poverty and disadvantage (Hulse et al. 2011). Drawing on the study of the six disadvantaged localities, the chapter considers the relevance of these arguments to the Australian urban context and generates some tentative policy prescriptions on how spatial concentrations of disadvantage in Australian cities might be addressed more effectively.

4.2

Conceptualising policy responses to spatial disadvantage

As outlined in Chapter 2, spatial disadvantage is often conceived in three overlapping ways: locational disadvantage; the concentration of social problems within specific neighbourhoods; and concentrated disadvantage (involving a disproportionate number of people conceived as being socio-economically disadvantaged living in particular places). In some areas, these forms of disadvantage converge, such that low-income groups find themselves ‘stuck’ in places that compound their disadvantage further (van Gent et al. 2009). Our analysis of these processes in the AHURI program of research, however, demonstrates that ‘getting stuck’ in places that compound disadvantage is by no means inevitable and that some areas classified as disadvantaged can still be relatively well-connected, rich in local services and dynamic in terms of residential mobility, reflecting an elevated level of optimism and aspiration among at least a proportion of residents (see Hulse et al. 2014; Cheshire et al. 2014). Nevertheless, an awareness of these different forms of disadvantage has informed the design of policies aimed at addressing them. In simple terms, scholars often make reference to policies being either people-focused in that they seek to improve the welfare of low-income groups regardless of where they live, or place-focused which tackle spatial disadvantage in designated localities (Bolton 1992). A third way, which has emerged in recent decades, takes the form of integrated area-based initiatives (ABIs). As recently exemplified in the UK’s New Deal for Communities (Lawless 2006), these simultaneously seek to address people and place-based forms of disadvantage through holistic strategies of neighbourhood renewal and socio-economic improvement of the local population. The reality of spatially concentrated disadvantage, however, means that it is often much more difficult to disentangle the people and place dimensions of specific interventions given that attempts to change a neighbourhood are often driven by the goal of improving the lives of the local population (e.g. through housing upgrades), while strategies to improve the socio-economic circumstances of disadvantaged people are equally likely to impact upon places where such populations are concentrated. The framework developed by Griggs et al. (2008) in Figure 1 below helps to clarify these points of connection and distinction by differentiating between the foci of policies (whether they are focused on people or places) and their intended impacts (upon people or places). From this, they identify five types of policy interventions depending on the relative importance they attach to people and places in their foci and impacts.

48

Figure 1: Policy objectives and targeting in relation to person and place, with examples

 Policy type 1: place-focused/place impact: strategic planning or city growth policies; land and

infrastructure development initiatives; or place improvement initiatives which may benefit local residents but which are implemented with relatively little attention to the effects upon them.  Policy type 2: place-focused/people impact: similarly focused on local area improvements but

with the explicit aim of improving the lives of both existing and future residents. Examples include physical renewal or upgrades of social housing, policies or social or tenure mix and community safety strategies to address crime and anti-social behaviour.  Policy type 3: people-focused/place impact: these target specific groups in a local area to

improve quality of life for those groups and to impact positively on the locality as a whole. They focus on ‘people in place’ through programs such as local employment initiatives, youth engagement strategies and support for new migrants.  Policy type 4: people-focused/people impact: these address individual and welfare issues but

do so without consideration of local circumstances. They include mainstream social welfare and economic policies such as unemployment or disability pensions and rental assistance schemes.  Policy type 5: people and place-focused/people and place impact: these arise in more limited

cases in the form of integrated area-based interventions or neighbourhood renewal. They often involve physical renewal combined with socio-economic interventions, capacity building, improved coordination of services, partnerships, and enhanced community participation. What this framework illustrates is that people- and place-based policies are commonly interwoven in their focus and impacts, even if this is often not explicitly articulated in their aims. In what follows, the kinds of policy interventions identified in the six study localities are classified according to this schema, with particular attention directed at those that are oriented to both places and people, either by focusing on one but impacting upon the other (policy types 2 and 3) or by explicitly seeking to tackle both (policy type 5). These three policy types are depicted in the shaded area in Figure 1 above. Policy type 1 is also briefly considered given that regional/city planning activities and place-improvement strategies may have (unintended and sometimes detrimental) consequences for local populations. Policy type 4 has no specific orientation to place, but questions continue to be raised about whether such policies remain the most appropriate for addressing disadvantage on a broad scale. If so, then it suggests, at the very least, that the

49

spatial impacts of these otherwise ‘spatially blind’ policies should be better understood and made more explicit.

4.3

The anatomy of spatial disadvantage in Australia: comparing six localities

One of the core findings of this AHURI series has been the diversity with which spatially concentrated disadvantage is manifested in large Australian cities. The disadvantaged suburbs typology presented in Table 5 in Chapter 3 demonstrates this clearly by showing how suburbs can be categorised into four types based on their key distinguishing socio-economic and housing market features. These were: Isolate suburbs; Lower-priced suburbs; Marginal suburbs; and Improver suburbs. Understanding the nature and extent of this diversity—as well as features that are common to all—is a necessary first step in assessing policy responses to urban disadvantage since one cannot determine the spatial sensitivity of urban social policy without first having a clearer picture of the anatomy of disadvantage that these policies seek to address. The experiences and forms of disadvantage found across the cities under study have been extensively documented in earlier reports produced for this series (see e.g. Hulse et al. 2014; Cheshire et al. 2014) as well as in Chapter 3. For the purpose of this chapter, tables from these reports relating to the demographic composition of the six study localities (including the educational and employment characteristics of the resident population), and the housing market characteristics of these and other disadvantaged areas relative to their respective Greater Metropolitan Area (GMA), been reproduced in Appendix 2 in Tables A1–A4. The key points of significance from these tables are highlighted here as context for the analysis of policy interventions that follow. Table 11: Summary of the nature of place- and people-based disadvantage in different localities People-based forms of disadvantage

Place-based forms of disadvantage

Isolate suburbs Emerton/Mount Druitt (Syd)

Comparable levels of Australian-born to GMA, but higher Aboriginal and Torres Strait Islanders. High percentage of loneparent families. Very high unemployment including youth unemployment. Low employment participation rates; low skilled/status jobs and very low incomes

Stigmatised, limited job opportunities; house prices and rents well below city medians. Poor access to transport, facilities and services; and perceived issues with crime and safety. Concentration of social rental (26.5%). Distant from city core.

Lower priced suburbs Auburn (Syd) Springvale (Melb)

High recent overseas arrivals (incl. refugees) and low Australian born; low Aboriginal and Torres Straight Islanders. Higher than GMA on non-traditional family households. Participation rate lower than GMA and higher levels of unemployment but youth unemployment close to city averages. Higher than city for low skilled/low status jobs.

Limited job opportunities despite transport connectivity to other centres. Affordable entry markets for sales but rents more than 85 per cent of city medians. Perceptions of crime (Auburn). Distant from the city core.

Marginal suburbs Russell Island (Bris)

High on older, single-person households needing assistance with core activities. Similar rate of Australian-born as GMA. Much lower rate of employment participation and very high rates of unemployment and youth unemployment.

Stigmatised. Remote from labour markets. High on home ownership and rental is 26.5 per cent. House prices much lower than city medians and rents just over half of city medians. Disconnected from the city more broadly. Poorly serviced and poor access to facilities and services. Very limited employment.

Improver suburbs Braybrook (Melb) Logan Central (Bris)

Mixed population: moderate new arrivals (incl. refugees); longer-term migrants; higher Aboriginal and Torres Strait Islanders. Higher on lone-parent families and ‘other’ family types, higher unemployment than GMA.

Stigmatised, limited employment, sales prices moving toward city medians and rents more than 85 per cent of city medians (Braybrook although not Logan). Perceptions of crime. Concentrations of social rental (20.1% Braybrook & 16.5% Logan Central). Close to city core.

50

Table 11 above provides a convenient summary of the different ways that people and place-based disadvantage manifest themselves in the four locality types and the six empirical locality areas. One of the most noticeable features is the variability in the way people-based disadvantage is manifested, such that each locality type is home to a distinct demographic. Where Isolate suburbs (e.g. Emerton/Mount Druitt) have levels of Australian-born residents that are comparable to their respective GMA but have higher proportions of Aboriginal and Torres Strait Islander people, Lower priced suburbs (e.g. Auburn and Springvale) have much higher proportions of overseas arrivals (including refugees). The population of Improver suburbs (e.g. Braybrook and Logan Central) is generally more mixed, with moderate proportions of new arrivals but also the presence of longer-term migrants who have subsequently settled in the area. Marginal suburbs (e.g. Russell Island), by contrast, have significantly higher proportions of older, single households who are likely to need assistance. As has been shown in Chapter 3, in terms of their housing markets, the profiles of the localities also support the proposition that disadvantage is now cross-tenure and no longer restricted to groups residing in social housing. While Isolate and Improver suburbs both contain large tracts of social housing, even here, it is not the dominant tenure. Both Lower priced and Marginal suburbs have housing markets largely comprised of low cost private rental and lower value home ownership. Regardless of any variation in housing markets and composite population, however, it is notable that the one feature that all disadvantaged areas share in common is high or very high levels of unemployment, ranging from 89 per cent to 227 per cent above their city averages (see Table A2 for more detail). This is indicative of the strong correlation between disadvantage and unemployment for those who are marginal to the job market, such as refugees, young people without a trade or Year 12 qualification, those with long-term disabilities, the elderly and single parents (Pacione 2004). Added to these groups are those in insecure low wage employment, especially those who are employed in the ‘shadow’ cash economy—aptly termed the 'precariat’ by Standing (2012). In addition, though, it also raises questions to be answered later about the effectiveness of place-based interventions in addressing one of the most fundamental features of people-based disadvantage. Turning now to the forms of spatial disadvantage encountered within disadvantaged areas, the constellations of disadvantaging features found in each of the case study locality types were described in detail in Chapters 4, 5 and 6 of Cheshire et al. (2014). Table 12 below provides a convenient summary of these findings. The darker red shading indicates where these features are thought to have a pronounced impact upon the people living there, while the lighter yellow shading points to those that are less profound or inconclusive. Reading across the rows of the table shows the nature and extent of place-based disadvantage in each locality type where, perhaps not surprisingly, it is the Marginal suburbs that appear most disadvantaged. Conversely, Springvale, as a Lower-priced suburb, appears the least disadvantaged in place terms, only featuring limited employment prospects and relatively low levels of affordable housing. Reading the table columns downwards allows for comparison among locality types to identify points of variation or commonality. Encouragingly, few of the localities appear to suffer from low levels of attachment among the local population, suggesting that important social-interactive processes such as social cohesion, collective efficacy and social capital are not inevitably diminished in disadvantaged areas. Similarly, few of the suburbs could be defined as poorly serviced, possibly because the concentration of disadvantaged populations in those areas has prompted the establishment of local services and organisations that cater for their needs. Only the Isolate (Emerton/Mount Druitt) and the Marginal suburbs (Russell Island) can be described as having poor connectivity to other areas of the city, and (potentially along with Logan Central) do not appear ‘dynamic’ in the way that the other suburbs do by virtue of their population growth, active property and labour markets and general air of vibrancy. Despite this, case study evidence suggests that there are a number of common experiences that run across most, if not all, disadvantaged locality types. As summarised in Chapter 3, all but

51

Springvale (which was thought by residents to have shed its negative reputation) appear to have been bestowed with a stigmatised identity that both offends, and reportedly disadvantages, residents, particularly at the point of trying to secure employment. Additionally, all have poor employment prospects, despite some suburbs such as Auburn and Springvale being local commercial and economic hubs. In some cases, the limited employment prospects are the outcome of restructured local labour markets brought on by processes such as deindustrialisation, but they are also the result of the increasingly centralised pattern of economic activity in Australian cities as described in Chapter 3. Finally, and perhaps counter-intuitively, despite disadvantaged suburbs having low property costs (rents and prices) relative to the rest of the city, limited housing affordability and the presence of housing stress appear to be persistent features of disadvantaged areas, most obviously because household incomes are typically low. But, as Hulse et al. (2014) have also reported, and as Table A4 illustrates, there has been a contraction of low-cost housing in disadvantaged areas over the last decade as housing prices in those areas start to ‘catch up’ with city averages. In summarising the ways that places may disadvantage their residents—and potentially compound existing forms of disadvantage brought on by the poor socio-economic circumstances of certain groups among them—the dominant ‘drivers’ of this outcome appear to be a combination of a stigmatised neighbourhood identity, limited employment opportunities and poor housing affordability. Arguably, the former is a behavioural outcome of the way the latter two ‘structural’ processes underpin the relative socio-economic disadvantages that have become associated with these places. The extent to which these limiting features of place have, or can be, addressed through policy interventions is explored in the next section.

52

Table 12: Dimensions of place disadvantage by suburb type—as exemplified by case study suburbs Dimensions of place disadvantage Suburb type/case study suburb(s)

Location

Concentrations of social housing

Low emp. prospects

Poor access

Poorly serviced

limited affordable 1 housing

High (fear of) crime

low attachment

Stigma

Dynamic?

far outer

Y

Y

Y

N

Y 21%

Y

N

YY

N

1. Isolate Emerton 2. Lower priced Auburn

middle

Y

N

N

Y 33%

Y

N

Y

Y

Springvale

outer

Y

N

N

Y 19.5%

N

N

N

Y

far outer

Y

Y

Y

Y 26.7%

N

?

YY

N

3. Marginal Russell Island 4. Improver Braybrook

inner

Y

Y

N

N

Y 20%

Y

N

YY

Y

Logan Central

outer

Y

Y

N

N

Y 33%

Y

N

YY

?

1

Note: Percentage figures indicate proportion of housing stress based on % of low-income h/holds with weekly h/hold income of